Return-Path: <@cs.wustl.edu:sandholm@siren.cs.wustl.edu> Received: from cs.wustl.edu by paris.ics.uci.edu id aa02003; 11 Nov 97 16:33 PST Received: from siren.cs.wustl.edu (sandholm@siren.cs.wustl.edu [128.252.165.132]) by rainier.cs.wustl.edu (8.8.5/CTS-JEK1.1) with ESMTP id SAA24124; Tue, 11 Nov 1997 18:33:21 -0600 (CST) Received: (from sandholm@localhost) by siren.cs.wustl.edu (8.8.5/CTS-JEK1.1) id SAA27131; Tue, 11 Nov 1997 18:33:19 -0600 (CST) Date: Tue, 11 Nov 1997 18:33:19 -0600 (CST) Message-Id: <199711120033.SAA27131@siren.cs.wustl.edu> From: Tuomas Sandholm To: smyth@sifnos.ics.uci.edu, sandholm@siren.cs.wustl.edu Subject: AAAI-98 tutorial proposal: Economically Founded Multiagent Systems Hi Padhraic, Please acknowledge receipt of this email. It was a pleasure to meet you at the DARPA Young Investigator Workshop this summer! I would like to propose a tutorial for AAAI-98. Let me know what you think. The proposed tutorial is an updated version of the tutorials that I have given at Agents-97, IJCAI-97, and at several industrial sites, including HP, Mitsubishi, and BusinessBots. In my opinion it has been successful every time it has been given so far: it drew about 31 people at Agents-97, and about 24 people at IJCAI-97. All the participants that I have talked to so far have found it useful and interesting. As a case in point, Yoav Shoham, one of the other experts in economically founded multiagent systems (MAS), took my tutorial at IJCAI-97 and said it was the best one he had ever heard on these topics (and he has heard most other MAS researchers' tutorials). He might be willing to comment on it to you if you would like. The evaluation forms at Agents-97 averaged about 4.5 on a five point scale (Joerg Muller can comment on this if you would like). I have invested enormous effort into preparing the tutorial. Just preparing the slides took 30 full-time equivalent days of work. Of course the material also reflects my 7 years of studying MAS, microeconomics, and game theory, as well as my experience in building and analyzing such systems. I have given a full semester graduate course on these topics (multiagent systems, particularly economically founded ones) at Washington University. I have also given a full semester graduate course on resource-bounded reasoning (intelligent real-time systems) at Washington University, and will draw from this material when discussing the issues of bounded rationality in multiagent systems. One of the main things I would like to emphasize in this proposal is that the tutorial is not focused on my own work (as often happens in tutorials), but is a true broad overview of the field. I propose the title "Economically Founded Multiagent Systems" but other alternative titles would be good also: - Agents in Electronic Markets - Agent-Mediated Electronic Trading - Automated Negotiation - Automated Negotiation and Coalition Formation - Microeconomic Algorithms for Computational Multiagent Systems The rest of this message contains more detailed items regarding the tutorial. 1. A detailed outline of the tutorial. Economically Founded Multiagent Systems A. Introduction to multiagent systems consisting of self-interested agents 1. What are multiagent systems and automated negotiation systems? 2. Why do we need/want them? 3. Autonomous, distributed agents 4. Open systems, no centralized designer 5. Self-interested vs. cooperative agents 6. Normative vs. non-normative methods in MAS 7. Designing interaction protocols 8. Growing importance - Technology push: x Internet, WWW, EDI, KQML, Java, Telescript, Concordia, Odyssey, Voyager, Aglets, ... x Collaboration technology - Application pull: x Electronic commerce of goods, info, bandwidth, processing, and storage x Virtual enterprises and agile manufacturing x Coordination of distributed operations x Other trends contributing to application pull 9. Relationships to other disciplines 10. Example applications - Multienterprise agile manufacturing planning & scheduling - Internet commerce, e.g. auctionbots - Electricity markets (Cases: California, New England, Finland, Sweden) - Heating markets for building environments - Multiagent information gathering on the web - Collective article rating on the web - Computer networks x Bandwidth allocation x Video on demand x Mirror site allocation x Dynamic pricing x Provider (and/or user) collusion - OS & mobile agents: allocation of storage & computation - Vehicle routing among independent dispatch centers - Treatment scheduling across hospitals - Electronic trading (Case: NASDAQ) - Multicontractor software engineering - Multirobot systems... 11. Agenthood - Utility theory - Full vs. bounded rationality 12. Distributed vs. centralized - Naturality - Reliability - Adaptability - Development & management - Efficiency (a critical inquiry) - Bandwidth bottleneck (a critical inquiry) 13. Evaluation criteria for multiagent systems 1. Computational efficiency 2. Distribution of computation 3. Communication efficiency 4. Social welfare 5. Pareto efficiency 6. Individual rationality 7. Stability 8. Symmetry 9. Others B. Game theoretic analysis tools 1. Terminology 2. Solution concepts - Dominant strategy equilibrium (Case: Prisoner's Dilemma game) - Nash equilibrium (Case: Battle of the Sexes Game) x Criticisms of Nash equilibrium (nonexistence, nonuniqueness, computability) x Existence theorems - Refinements of Nash equilibrium 3. The Revelation Principle and why it does not hold among computational agents C. Voting 1. Voting setting 2. Truthful voting - Undesirable properties of common voting protocols x Agenda paradox x Pareto dominated winner paradox x Inverted order paradox x Majority winner paradox - Arrow's impossibility theorem 3. Strategic voting - Gibbard-Satterthwaite impossibility theorem - Ways to get around the impossibility x Restricted domains (Case: Clarke tax mechanism for truth extraction in quasilinear environments; Applications in multiagent planning) x Randomization (Case: Hat protocol) x Complexity D. Auctions 1. Auction settings - Private value auctions - Common value auctions - Hybrids 2. Auction protocols and their properties - All-pay auctions - Ascending (English) auction - First-price sealed-bid auction - Descending (Dutch) auction - Second-price sealed-bid (Vickrey) auction 3. Results for private value auctions - Optimal strategies - Allocation efficiency - Revenue Equivalence Theorem - Revenue nonequivalence with risk averse bidders/auctioneer 4. Results for non-private value auctions - Optimal strategies - Allocation efficiency - Winner's curse - Revenue nonequivalence - Settings with asymmetric information among agents 5. Collusion in auctions (Case: Self-enforcing collusive agreements) 6. Vulnerability to shills 7. Vulnerability to a lying auctioneer (Case: Third party auctionbots) 8. Auctioneer's other possibilities - Bidding - Minimum prices - Refusing to sell 9. Undesirable private information revelation 10. Untruthful bidding with local uncertainty in the Vickrey auction 11. Wasteful counterspeculation even in the Vickrey auction - Costly computation actions - Costly information gathering actions 12. Sequential interrelated auctions - Complexity and local optima (Case 1: FCC bandwidth auctions; Case 2: Dynamic pricing in computer networks) - Inefficient allocation under truthful myopic bidding - Lying in interrelated auctions 13. Continuous double auction protocol (Case: New York Stock Exchange) 14. Existing auction-based CS applications - Network bandwidth allocation - Computation allocation E. General equilibrium based market mechanisms 1. Example CS applications - Power load management (Case: Swedish electricity market) - Distributed design - Mirror site allocation - Network flow routing 2. General equilibrium market setting - Consumers and their choice sets - Producers and their choice sets - Commodities and assumptions regarding them - Prices 3. Definition of a general equilibrium 4. Properties of a general equilibrium - Pareto efficiency: First Fundamental Theorem of Welfare Economics - Second Fundamental Theorem of Welfare Economics - Existence conditions for a general equilibrium - Uniqueness conditions for a general equilibrium 5. Limitations of the basic general equilibrium model - Price taking (competitive) assumption x Sandholm&Ygge speculation scheme and its convergence - Divisible vs. discrete goods - Convex production possibilities sets - Externalities - Universal prices - No transfers before convergence - Single-shot 6. Algorithms for reaching a general equilibrium - Price tatonnement x Convergence conditions x Conceptual problems - Variable step (Newtonian) price tatonnement x Convergence conditions - WALRAS [Wellman et al] x Convergence conditions - Quantity tatonnement x Convergence conditions x Anytime property 7. Pros of general equilibrium methods 8. Cons of general equilibrium methods F. Coalition formation 1. Coalition formation as - Desirable (Case: Vehicle routing among independent dispatch centers) - Undesirable (Case: Auctions) 2. Activities of coalition formation - Coalition structure generation - Optimization - Payoff division 3. Approach of classic game theory 4. High-level classification of coalition games - Domain classification for rational agents - Domain classification for bounded rational agents 5. Solution concepts - Core (Case: Coalition formation among ATM network providers) - Shapley value - Coalition-proof Nash equilibrium - Strong Nash equilibrium 6. Core and Walrasian equilibrium 7. Research on reducing complexity of coalition formation - In the coalition structure generation activity (Case 1: Shehory&Kraus algorithm; Case 2: Ketchpel algorithm) - In the optimization activity (Case: Sandholm&Lesser method) - In the payoff division activity (Case 1: Transfer schemes (example runs are presented)); Case 2: Zlotkin&Rosenschein cryptographic algorithm) G. Contract nets 1. The contract net concept for task allocation 2. Desirability among computationally limited agents 3. Scaling up (Case: Vehicle routing among independent dispatch centers) 4. Marginal costs in automated contracting 5. Bidding and awarding while old bids are pending 6. Modern combinatorial contract types - Cluster contracts - Swap contracts - Multiagent contracts - OCSM contracts - Results x Necessity for optimal task allocation x Sufficiency for optimal task allocation x Anytime property 7. Leveled commitment contracts - Definition - Motivation - Uses - Comparison to contingency contracts - Insincere decommitting - Example using noncooperative equilibrium analysis x Contract enabling theorem x Efficiency improvement theorem - Biased information 8. Hybrid multiagent search algorithms 9. Message congestion and agent saturation (Case: TRACONET experience of avoiding these problems in an asynchronous distributed implementation) 10. Tradeoffs resulting from limited computation 2. A brief description of the tutorial, suitable for inclusion in the registration brochure. Economically Founded Multiagent Systems Prof. Tuomas Sandholm Department of Computer Science Washington University St. Louis, MO, 63130 In multiagent systems---e.g. for agent-mediated electronic commerce---computational agents make contracts on behalf of the real world parties that they represent. Such automated negotiation saves human negotiation time, and computational agents are often better at finding beneficial deals than humans are in combinatorially and strategically complex settings. The significance of automated multiagent systems is increasing due to the developing communication infrastructure, the advent of electronic commerce, the industrial trend toward outsourcing, and the growing need to coordinate distributed operations. Important applications include electronic commerce and trading, manufacturing planning and scheduling among multiple agile enterprises, electricity markets, allocating and pricing bandwidth in multi-provider multi-consumer computer networks, network management, multiagent information gathering on the web, digital libraries, distributed vehicle routing among independent dispatch centers, resource allocation in distributed operating systems, meeting scheduling, scheduling of patient treatments across hospitals, classroom scheduling, and planning and scheduling of multi-contractor software projects, to name just a few. A key research goal is to design open distributed systems in a principled way that leads to globally desirable outcomes even though every participating agent only considers its own good and may act insincerely. The tutorial covers relevant results in AI, game theory, market mechanisms, voting, auctions, coalition formation, and contract nets. Emphasis is given to rigorous concepts, results, and algorithms - both classic ones from microeconomics and recent ones from the distributed AI community - that have direct applications to computational multiagent systems. Implementation experiences will be shared. Effects of different computational limitations (agents' bounded rationality) are discussed as a key feature that has not received adequate attention. Examples of real-world applications will be presented. Dr. Tuomas Sandholm is assistant professor of computer science at Washington University. He received the M.S. (B.S. included) with distinction in Industrial Engineering and Management Science from the Helsinki University of Technology, Finland, in 1991. From 1988 to 1992, he worked as a research scientist in the software industry. He earned the M.S. and Ph.D. degrees in computer science from the University of Massachusetts at Amherst in 1994 and 1996 respectively. He has published over 50 technical papers in forums such as AI Journal, IJCAI, and AAAI. He has been a program committee member for thirteen major conferences, and a reviewer for nine journals and numerous conferences. He has seven years of experience designing efficient multiagent systems. This work has focused both on theory and implementations. He has also been involved in developing two fielded AI systems: a pension law expert system and a large-scale transportation optimization application. 3. A clear statement of the necessary background and of the potential target audience. The tutorial is targeted to the builder of (open) multiagent systems that consist of multiple self-interested agents. It also serves to make newcomers and executive level participants familiar with the issues in multiagent systems. The tutorial reviews work from microeconomics that has been found most relevant to computational agents. It also presents recent results from the multiagent systems (DAI) research community. Emphasis is given to results that have direct applications to computational multiagent systems. Applications themselves will be discussed. No specific background is required in economics or multiagent systems. A general familiarity with computer science is helpful, but even a layperson will be able to appreciate the basic issues raised and answered in this tutorial. The target audience includes: - Builders of open multiagent systems. - Builders of electronic market places (auctionbots, automated negotiation software, electronic stock trading protocols and monitoring daemons, etc.). - Researchers in MAS who know some microeconomic theory, but do not know the big picture of it well. - Researchers in MAS/DAI that do not know game theory, but would like to get a thorough and broad overview of how it can be used in computational MAS. - Non-MAS participants who are interested in knowing what questions the MAS community is interested in, and how they are tackling these questions. I believe that the tutorial would do the MAS community proud. - Executive/management level participants that want to know what the key issues are, and what MAS technologies they can use in their applications. - Electronic commerce people who are interested in how agents can beneficially mediate electronic commerce. - Press who want to cover MAS. 4. A description of why the tutorial topic is of interest to a substantial part of the conference audience The techniques of this tutorial are not well known in the "agents" or even in the "multiagent systems" community. At the same time they are absolutely crucial for coordinating multiple self-interested agents that represent different real world parties. Recent developments in technology and business practices have led to ubiquity of such settings. These application domains include electronic commerce and trading, manufacturing planning and scheduling among multiple agile enterprises, markets for electricity, allocating and pricing bandwidth in multi-provider multi-consumer computer networks, network management, multiagent information gathering on the web, distributed vehicle routing among independent dispatch centers, resource allocation in distributed operating systems, meeting scheduling, scheduling of patient treatments across hospitals, classroom scheduling, and planning and scheduling of multi-contractor software projects, to name just a few. In such settings, robustly (nonmanipulatively) coordinated multiagent systems can save users' time, but they may also achieve better solutions (e.g. by enhanced negotiation and coalition formation) than human agents can in combinatorially and strategically complex domains. For the general AI, agents, or alife researcher, it is very important and probably interesting to see what are some of the specific issues that "multiagent" researchers work on. At the same time, the tutorial serves the multiagent researcher who may not be familiar with microeconomic foundations of the field, or who may know some of the microeconomics, but not the big picture of these results. The focus on electronic commerce is likely to draw a large audience since ecommerce is such a booming industry, and agent mediation in it is understood to be important - yet the technologies that implement such mediation are not well understood by the public. 5. A brief presentation of the speaker(s), including name, postal address, phone and fax numbers, email address, background in the tutorial area (e.g. teaching experience, examples of work in the area). A postscript of my CV is annexed. 6. The proposed duration of the tutorial Preferably full day since I have the slides ready for a full day version. If necessary, I can give a half day version, but of course a shorter version will not be as complete. I have given the full day version 3 times, and the half day version 2 times. 7. An estimate of the volume of printed material to be provided to each participant 141 pages. Two pages can be printed on the same sheet to reduce copying to 71 sheets per participant. I will mail you color copies of the slides tomorrow for your information. Best regards, Tuomas *-----------------------------------------------------------------------* | Tuomas Sandholm, Assistant Professor Voice: (314) 935-4749 | | Department of Computer Science, Campus Box 1045 Fax: (314) 935-7302 | | Washington University | | One Brookings Drive http://www.cs.wustl.edu/~sandholm | | St. Louis, MO 63130-4899 sandholm@cs.wustl.edu | *-----------------------------------------------------------------------* --------- postscript of CV follows ----------- %!PS-Adobe-2.0 %%Creator: dvipsk 5.58f Copyright 1986, 1994 Radical Eye Software %%Title: cv.dvi %%Pages: 29 %%PageOrder: Ascend %%BoundingBox: 0 0 612 792 %%EndComments %DVIPSCommandLine: dvips -o cv.ps cv.dvi %DVIPSParameters: dpi=300, compressed, comments removed %DVIPSSource: TeX output 1997.11.11:1822 %%BeginProcSet: texc.pro /TeXDict 250 dict def TeXDict begin /N{def}def /B{bind def}N /S{exch}N /X{S N}B /TR{translate}N /isls false N /vsize 11 72 mul N /hsize 8.5 72 mul N /landplus90{false}def /@rigin{isls{[0 landplus90{1 -1}{-1 1} ifelse 0 0 0]concat}if 72 Resolution div 72 VResolution div neg scale isls{landplus90{VResolution 72 div vsize mul 0 exch}{Resolution -72 div hsize mul 0}ifelse TR}if Resolution VResolution vsize -72 div 1 add mul TR[matrix currentmatrix{dup dup round sub abs 0.00001 lt{round}if} forall round exch round exch]setmatrix}N /@landscape{/isls true N}B /@manualfeed{statusdict /manualfeed true put}B /@copies{/#copies X}B /FMat[1 0 0 -1 0 0]N /FBB[0 0 0 0]N /nn 0 N /IE 0 N /ctr 0 N /df-tail{ /nn 8 dict N nn begin /FontType 3 N /FontMatrix fntrx N /FontBBox FBB N string /base X array /BitMaps X /BuildChar{CharBuilder}N /Encoding IE N end dup{/foo setfont}2 array copy cvx N load 0 nn put /ctr 0 N[}B /df{ /sf 1 N /fntrx FMat N df-tail}B /dfs{div /sf X /fntrx[sf 0 0 sf neg 0 0] N df-tail}B /E{pop nn dup definefont setfont}B /ch-width{ch-data dup length 5 sub get}B /ch-height{ch-data dup length 4 sub get}B /ch-xoff{ 128 ch-data dup length 3 sub get sub}B /ch-yoff{ch-data dup length 2 sub get 127 sub}B /ch-dx{ch-data dup length 1 sub get}B /ch-image{ch-data dup type /stringtype ne{ctr get /ctr ctr 1 add N}if}B /id 0 N /rw 0 N /rc 0 N /gp 0 N /cp 0 N /G 0 N /sf 0 N /CharBuilder{save 3 1 roll S dup /base get 2 index get S /BitMaps get S get /ch-data X pop /ctr 0 N ch-dx 0 ch-xoff ch-yoff ch-height sub ch-xoff ch-width add ch-yoff setcachedevice ch-width ch-height true[1 0 0 -1 -.1 ch-xoff sub ch-yoff .1 sub]/id ch-image N /rw ch-width 7 add 8 idiv string N /rc 0 N /gp 0 N /cp 0 N{rc 0 ne{rc 1 sub /rc X rw}{G}ifelse}imagemask restore}B /G{{id gp get /gp gp 1 add N dup 18 mod S 18 idiv pl S get exec}loop}B /adv{cp add /cp X}B /chg{rw cp id gp 4 index getinterval putinterval dup gp add /gp X adv}B /nd{/cp 0 N rw exit}B /lsh{rw cp 2 copy get dup 0 eq{pop 1}{ dup 255 eq{pop 254}{dup dup add 255 and S 1 and or}ifelse}ifelse put 1 adv}B /rsh{rw cp 2 copy get dup 0 eq{pop 128}{dup 255 eq{pop 127}{dup 2 idiv S 128 and or}ifelse}ifelse put 1 adv}B /clr{rw cp 2 index string putinterval adv}B /set{rw cp fillstr 0 4 index getinterval putinterval adv}B /fillstr 18 string 0 1 17{2 copy 255 put pop}for N /pl[{adv 1 chg} {adv 1 chg nd}{1 add chg}{1 add chg nd}{adv lsh}{adv lsh nd}{adv rsh}{ adv rsh nd}{1 add adv}{/rc X nd}{1 add set}{1 add clr}{adv 2 chg}{adv 2 chg nd}{pop nd}]dup{bind pop}forall N /D{/cc X dup type /stringtype ne{] }if nn /base get cc ctr put nn /BitMaps get S ctr S sf 1 ne{dup dup length 1 sub dup 2 index S get sf div put}if put /ctr ctr 1 add N}B /I{ cc 1 add D}B /bop{userdict /bop-hook known{bop-hook}if /SI save N @rigin 0 0 moveto /V matrix currentmatrix dup 1 get dup mul exch 0 get dup mul add .99 lt{/QV}{/RV}ifelse load def pop pop}N /eop{SI restore userdict /eop-hook known{eop-hook}if showpage}N /@start{userdict /start-hook known{start-hook}if pop /VResolution X /Resolution X 1000 div /DVImag X /IE 256 array N 0 1 255{IE S 1 string dup 0 3 index put cvn put}for 65781.76 div /vsize X 65781.76 div /hsize X}N /p{show}N /RMat[1 0 0 -1 0 0]N /BDot 260 string N /rulex 0 N /ruley 0 N /v{/ruley X /rulex X V}B /V {}B /RV statusdict begin /product where{pop product dup length 7 ge{0 7 getinterval dup(Display)eq exch 0 4 getinterval(NeXT)eq or}{pop false} ifelse}{false}ifelse end{{gsave TR -.1 .1 TR 1 1 scale rulex ruley false RMat{BDot}imagemask grestore}}{{gsave TR -.1 .1 TR rulex ruley scale 1 1 false RMat{BDot}imagemask grestore}}ifelse B /QV{gsave newpath transform round exch round exch itransform moveto rulex 0 rlineto 0 ruley neg rlineto rulex neg 0 rlineto fill grestore}B /a{moveto}B /delta 0 N /tail {dup /delta X 0 rmoveto}B /M{S p delta add tail}B /b{S p tail}B /c{-4 M} B /d{-3 M}B /e{-2 M}B /f{-1 M}B /g{0 M}B /h{1 M}B /i{2 M}B /j{3 M}B /k{ 4 M}B /w{0 rmoveto}B /l{p -4 w}B /m{p -3 w}B /n{p -2 w}B /o{p -1 w}B /q{ p 1 w}B /r{p 2 w}B /s{p 3 w}B /t{p 4 w}B /x{0 S rmoveto}B /y{3 2 roll p a}B /bos{/SS save N}B /eos{SS restore}B end %%EndProcSet TeXDict begin 40258431 52099146 1000 300 300 (cv.dvi) @start /Fa 1 16 df15 D E /Fb 68 123 df12 D34 D38 D<120C121E123FA2121D1202A31204A212081210122012401280080F75A20F>I<1480EB 010013025B5B5B13305B5BA2485A48C7FCA21206A2120E120C121C1218A212381230A212 70A21260A212E0A35AAD12401260A21220123012107E113278A414>I<13087F13061302 1303A27F1480AD1303A31400A25BA21306A2130E130CA2131C131813381330A25BA25B48 5AA248C7FC120612045A5A5A5A5A113280A414>I<120E121EA41202A21204A21208A212 10122012401280070F7D840F>44 DI<127012F8A212F012 E005057A840F>I<150815181530A2156015C0A2EC0180A2EC03001406A25CA25C5CA25C A25C495AA249C7FCA21306A25B5BA25BA25B5BA2485AA248C8FC1206A25AA25A5AA25AA2 5AA25A1D317FA419>II<13011303A21306131E132EEA03CEEA001CA41338A41370A413E0 A4EA01C0A4EA0380A41207EAFFFC10217AA019>III<14181438A21470 A314E0A314C01301148013031400A21306A25BA25B1310EB3180EB61C0EB438013831201 EA03033802070012041208EA3FC7EA403E38800FF038000E00A25BA45BA31330152B7EA0 19>III<380278023804FC041207000F1308EB 0C18381E0470381803A03830006000201340006013C000401380EA8001EA000314005B13 06130EA25BA2133C13381378A25BA3485AA312035BA26C5A172279A019>I57 D<1207EA0F80A21300120EC7FCAB127012F8A25A5A09157A940F >I<1403A25CA25CA25C142FA2EC4F80A21487A2EB01071302A21304A213081318131013 20A290387FFFC0EB40031380120113001202A25AA2120C003CEB07E0B4EB3FFC1E237DA2 24>65 D<90B512E090380F0038151C151E011E130E150FA349130E151EA2153C49137815 70EC01E0EC078090B5FC9038F001E0EC00F01578485A1538153CA248481378A315F0485A EC01E0EC03C0EC0700380F001EB512F020227DA122>I<027F138090390380810090380E 00630138132749131F49130E485A485A48C7FC481404120E121E5A5D4891C7FCA35AA55A 1520A25DA26C5C12704AC7FC6C130200185B001C5B00061330380381C0D800FEC8FC2124 79A223>I<90B512F090380F003C150E81011EEB0380A2ED01C0A25B16E0A35BA449EB03 C0A44848EB0780A216005D4848130E5D153C153848485B5DEC03804AC7FC000F131CB512 F023227DA125>I<90B6128090380F00071501A2131EA21600A25BA2140192C7FCEB7802 A21406140EEBFFFCEBF00CA33801E008A21504EC0008485AA25DA248485B15605D140138 0F0007B65A21227DA121>I<90B6FC90380F000F1503A2131EA21502A25BA214011500EB 7802A21406140EEBFFFCEBF00CA33801E008A391C7FC485AA4485AA4120FEAFFF820227D A120>I<027F138090390380810090380E00630138132749131F49130E485A485A48C7FC 481404120E121E5A5D4891C7FCA35AA4EC3FFC48EB01E0A34A5AA27E12704A5A7E001813 0F001C131300060123C7FC380381C1D800FEC8FC212479A226>I<9039FFF87FFC90390F 000780A3011EEB0F00A449131EA4495BA490B512F89038F00078A348485BA44848485AA4 4848485AA4000F130739FFF07FF826227DA124>II<903807FFC09038003C00A35CA45C A4495AA4495AA4495AA449C7FCA212381278EAF81EA2485AEA40385BEA21E0EA1F801A23 7CA11A>I76 DI<01FFEB0FFC90390F8001E01680EC C0000113EB0100A2EB11E0A201211302EB20F0A39038407804A3143C01805B143E141EA2 3901001F10140FA2EC0790000214A0A2EC03E0A2485C1401A2120C001E6D5AB47E26227D A124>I<14FE903807838090380C00E0013813704913385B4848131C485A48C7FC48141E 121E121C123CA25AA348143CA31578A25A15F0A2EC01E015C06C1303EC07800070140000 78130E00385B6C5B6C13E038070380D801FCC7FC1F2479A225>I<90B512E090380F0038 151E150E011E1307A449130FA3151E5B153C157815E09038F003C09038FFFE0001F0C7FC A2485AA4485AA4485AA4120FEAFFF020227DA121>I<14FE903807838090380C00E00138 13704913385B4848133C4848131C48C7FC48141E121EA25AA25AA348143CA3153815785A 15F0A2EC01E015C01403D8F01E13803970610700EB810E00385B001C5B000E13E0390783 80403801FD00EA0001EC808014811500EB0387EB01FEA25CEB00F01F2D79A225>I<90B5 12C090380F0070153C151C011E130EA2150FA249131EA3153C4913381570EC01E0EC0780 9038FFFC00EBF00E80EC0380D801E013C0A43903C00780A43907800F001501A2EC070212 0F39FFF0038CC812F020237DA124>I<903801F02090380E0C4090381802C0EB30011360 01E0138013C01201A200031400A291C7FCA27FEA01F813FF6C13E06D7EEB1FF8EB03FCEB 007C143C80A30020131CA3141800601338143000705B5C38C80180D8C607C7FCEA81FC1B 247DA21B>I<001FB512F8391E03C03800181418123038200780A200401410A2EB0F0012 80A200001400131EA45BA45BA45BA4485AA41203B5FC1D2277A123>I<393FFE03FF3903 C000781560152048481340A448C71280A4001EEB0100A4481302A4485BA400705B12F05C 12705C5C123038380180D81802C7FCEA0E0CEA03F0202377A124>I<3BFFE07FF00FF83B 1F000F8003C0001E913800018017001602A2001F5DA26C495BA2022F5BA291384F802002 475B14875EEB010701030181C7FC130201041382A201081384158C011013881590132001 A013A013C0D8078013C0A201005BA2000691C8FC140212042D2376A131>87 D<90397FF80FFC90390FC003C0D907801380ED02006E5A01035BECE018151001015B6E5A 01005B02F1C7FC14FA147CA3143CA2147E149EEB011EEB031FEB060F01047F1308EB1007 01207FEB40031380000180380300015A001F497E39FFC00FFE26227EA124>II<4813 8038020100EA0402485A485A485AA2485AA2485AEAB85CEAFC7EA2EAF87CEA7038110F73 A219>92 D97 DI<137EEA01C138030180EA0703EA0E07121C003C C7FC12381278A35AA45B12701302EA300CEA1830EA0FC011157B9416>I<143CEB03F8EB 0038A31470A414E0A4EB01C013F9EA0185EA0705380E0380A2121C123C383807001278A3 EAF00EA31410EB1C201270133C38305C40138C380F078016237BA219>I<13F8EA0384EA 0E02121C123C1238EA7804EAF018EAFFE0EAF000A25AA41302A2EA6004EA7018EA3060EA 0F800F157A9416>I<143E144714CFEB018F1486EB0380A3EB0700A5130EEBFFF0EB0E00 A35BA55BA55BA55BA45B1201A2EA718012F100F3C7FC1262123C182D82A20F>II<13F0EA0F E01200A3485AA4485AA448C7FC131FEB2180EBC0C0380F00E0A2120EA2381C01C0A43838 0380A3EB070400701308130E1410130600E01320386003C016237DA219>I<13C0EA01E0 13C0A2C7FCA8121E12231243A25AA3120EA25AA35AA21340EA7080A3EA71001232121C0B 217BA00F>I<13F0EA07E01200A3485AA4485AA448C7FCEB01E0EB0210EB0470380E08F0 1310EB2060EB4000EA1D80001EC7FCEA1FC0EA1C70487EA27F142038703840A3EB188012 E038600F0014237DA216>107 DI<391C0F80F8392610C10C394760660639 87807807A2EB0070A2000EEBE00EA44848485AA3ED38202638038013401570168015303A 7007003100D83003131E23157B9428>I<38380F80384C30C0384E4060388E8070EA8F00 128EA24813E0A4383801C0A3EB03840070138814081307EB031012E0386001E016157B94 1B>I<137EEA01C338038180380701C0120E001C13E0123C12381278A338F003C0A21480 130700701300130E130CEA3018EA1870EA07C013157B9419>I<3801C1F0380262183804 741C3808780CEB700EA2141EEA00E0A43801C03CA3147838038070A2EBC0E0EBC1C03807 2380EB1E0090C7FCA2120EA45AA3EAFFC0171F7F9419>I114 D<13FCEA018338020080EA0401EA0C03140090C7FC120F13F0EA07FC6C7EEA003E130F7F 1270EAF006A2EAE004EA4008EA2030EA1FC011157D9414>I<13C01201A4EA0380A4EA07 00EAFFF8EA0700A2120EA45AA45AA31310EA7020A213401380EA3100121E0D1F7C9E10> I<001E1360002313E0EA4380EB81C01283EA8701A238070380120EA3381C0700A31408EB 0E101218121CEB1E20EA0C263807C3C015157B941A>I<381C0180382603C0EA47071303 EA8701EA8E00A2000E13805AA338380100A31302A25B5B1218EA0C30EA07C012157B9416 >I<001EEB60E00023EBE0F0384380E1EB81C000831470D887011330A23907038020120E A3391C070040A31580A2EC0100130F380C0B02380613843803E0F81C157B9420>I<3803 C1E0380462103808347038103CF0EA203814601400C65AA45BA314203861C04012F11480 38E2C100EA4462EA383C14157D9416>I<001E133000231370EA438014E01283EA8700A2 380701C0120EA3381C0380A4EB0700A35BEA0C3EEA03CEEA000EA25B1260EAF038133048 5AEA80C0EA4380003EC7FC141F7B9418>I<3801E0203803F0603807F8C038041F803808 01001302C65A5B5B5B5B5B48C7FC120248138038080100485AEA3F06EA61FEEA40FCEA80 7013157D9414>I E /Fc 72 128 df12 D<903907F80FF090397FFE7FFC3A01F80FF80ED803E0EBE01F484848485AD80F801380A2 020F131E160C93C7FCA4B8FCA23A0F800F801FB23B7FF07FF0FFE0A22B237FA22F>14 D<1238127C12FE12FFA2127F123B1203A31206A2120C121C12181270122008117CA210> 39 D<13181330136013C01201EA0380EA0700A2120E121E121C123CA35AA412F85AAB7E 1278A47EA3121C121E120E7EA2EA0380EA01C012001360133013180D317BA416>I<12C0 12607E7E121C7E7EA2EA038013C0120113E0A3EA00F0A413F81378AB13F813F0A4EA01E0 A313C012031380EA0700A2120E5A12185A5A5A0D317DA416>I<1238127C12FE12FFA212 7F123B1203A31206A2120C121C12181270122008117C8610>44 DI<1238127C12FEA3127C123807077C8610>I<14181438A21470A214E0A3EB01C0A2EB03 80A3EB0700A3130EA25BA35BA25BA35BA2485AA3485AA248C7FCA3120EA35AA25AA35AA2 5AA25A15317DA41C>I<13FE3807FFC0380F83E0381F01F0383E00F8A248137CA312FC14 7EAD007C137CA36C13F8A2381F01F0380F83E03807FFC03800FE0017207E9F1C>I<1318 1378EA01F812FFA21201B3A7387FFFE0A213207C9F1C>I I<13FE3807FFC0380F07E0381E03F0123FEB81F8A3EA1F0314F0120014E0EB07C0EB1F80 3801FE007F380007C0EB01F014F8EB00FCA2003C13FE127EB4FCA314FCEA7E01007813F8 381E07F0380FFFC03801FE0017207E9F1C>I<14E013011303A21307130F131FA2133713 7713E7EA01C71387EA03071207120E120C12181238127012E0B6FCA2380007E0A790B5FC A218207E9F1C>I<00301320383E01E0383FFFC0148014005B13F8EA33C00030C7FCA4EA 31FCEA37FF383E0FC0383807E0EA3003000013F0A214F8A21238127C12FEA200FC13F0A2 387007E0003013C0383C1F80380FFF00EA03F815207D9F1C>II<12601278387FFFFEA214FC14F8A214F038E0006014C038C00180EB 0300A2EA00065B131C131813381378A25BA31201A31203A76C5A17227DA11C>I<13FE38 03FFC0380703E0380E00F05A1478123C123E123F1380EBE0F0381FF9E0EBFFC06C13806C 13C06C13E04813F0381E7FF8383C1FFCEA7807EB01FEEAF000143E141EA36C131C007813 387E001F13F0380FFFC00001130017207E9F1C>II<1238127C12FEA3127C12381200A81238127C12FEA3127C123807167C9510>I<1470 A214F8A3497EA2497EA3EB067FA2010C7F143FA2496C7EA201307F140F01707FEB6007A2 01C07F90B5FC4880EB8001A2D803007F14004880000680A23AFFE007FFF8A225227EA12A >65 DIIII IIII<48B51280A2390003F800B3A41218127E12FFA3 5C387E07E0387C0FC0381FFF80D807FCC7FC192280A11D>IIIIIIIII< 3801FE023807FF86381F01FE383C007E007C131E0078130EA200F81306A27E1400B4FC13 E06CB4FC14C06C13F06C13F86C13FC000313FEEA003F1303EB007F143FA200C0131FA36C 131EA26C133C12FCB413F838C7FFE00080138018227DA11F>I<007FB61280A2397E03F8 0F00781407007014030060140100E015C0A200C01400A400001500B3A248B512F0A22222 7EA127>III< B53A0FFFF00FFEA2260FF00090C712E000076E14C0A26C6C90393F800180A26D14C00001 EE0300A26D496C5A0000026F1306A2017F9038C7F00CA2028113F8013F01835BA29139C3 01FC38011F153016FE90260FE6005BA202FEEBFFE06D486D5AA26D486D5AA36D486DC7FC A36D48130EA237227FA13A>I89 D97 DIII<13FE3807FF80380F87C0381E01E0003E13F0EA7C0014F812FCA2 B5FCA200FCC7FCA3127CA2127E003E13186C1330380FC0703803FFC0C6130015167E951A >II<3801FE0F3907FFBF80380F87C7381F03E7391E01E000003E7FA500 1E5BEA1F03380F87C0EBFF80D809FEC7FC0018C8FCA2121C381FFFE06C13F86C13FE001F 7F383C003F48EB0F80481307A40078EB0F006C131E001F137C6CB45A000113C019217F95 1C>II<121C123E127FA3123E121CC7FCA7B4FCA2121FB2EAFFE0A20B247EA310>I<13 38137C13FEA3137C13381300A7EA03FEA2EA003EB3A5127812FC133C137CEA78F8EA7FE0 EA1F800F2E83A311>III<3AFF07F007F090391FFC1FFC3A1F303E303E0140 1340496C487EA201001300AE3BFFE0FFE0FFE0A22B167E9530>I<38FF07E0EB1FF8381F 307CEB403CEB803EA21300AE39FFE1FFC0A21A167E951F>I<13FE3807FFC0380F83E038 1E00F0003E13F848137CA300FC137EA7007C137CA26C13F8381F01F0380F83E03807FFC0 3800FE0017167E951C>I<38FF0FE0EB3FF8381FE07CEB803E497E1580A2EC0FC0A8EC1F 80A29038803F00EBC03EEBE0FCEB3FF8EB0FC090C8FCA8EAFFE0A21A207E951F>IIII<487EA41203A21207A2120F123FB5FCA2EA0F80ABEB 8180A5EB8300EA07C3EA03FEEA00F811207F9F16>I<38FF01FEA2381F003EAF147E14FE 380F81BE3907FF3FC0EA01FC1A167E951F>I<39FFE01FE0A2391F800700000F1306EBC0 0E0007130C13E000035BA26C6C5AA26C6C5AA2EB7CC0A2137F6D5AA26DC7FCA2130EA21B 167F951E>I<3AFFE3FF07F8A23A1F007800C09038807C01000F1580A23A07C07E030014 DE5D3903E1DF06148FD801F1138CEBF307A2D800FF13D8EBFE0315F890387C01F0A2013C 5BEB3800A225167F9528>I<39FFE07FC0A2390F801C006C6C5A6C6C5AEBF0606C6C5A38 00F980137F6DC7FC7F80497E1337EB63E0EBC1F03801C0F848487E3807007E000E133E39 FF80FFE0A21B167F951E>I<39FFE01FE0A2391F800700000F1306EBC00E0007130C13E0 00035BA26C6C5AA26C6C5AA2EB7CC0A2137F6D5AA26DC7FCA2130EA2130CA25B1278EAFC 3813305BEA69C0EA7F80001FC8FC1B207F951E>I<387FFFF0A2387C03E0387007C0EA60 0F38E01F8000C01300133E137EC65A5B485A00031330EA07E013C0380F8070121F383F00 60003E13E0EA7C03B5FCA214167E9519>I124 D127 D E /Fd 15 120 df<127012F8A3127005057A8410>46 D64 D<1304130EA3131FA2EB2F801327A2EB43C0A2EBC3E01381A248C67EA2487F13FF380200 78487FA3487F1218003C131F00FEEB7FE01B1A7F991F>97 D99 DII<39FFC1FF80391E003C00AB381FFFFC381E003CAC39FFC1FF80191A7E991F> 104 D108 D<00FEEB01FE001E14F0A200171302A238138004A33811C008A23810E010A3EB7020A3EB 3840A2EB1C80A3EB0F00A21306123800FEEB07FE1F1A7E9925>I<00FEEB3F80001FEB0E 00EB80041217EA13C0EA11E013F012101378137C133C131E130F14841307EB03C4EB01E4 A2EB00F4147CA2143C141C140C123800FE1304191A7E991F>II115 D<007FB5FC38701E0700601301124000 C0148000801300A300001400B0133F3803FFF0191A7F991D>I<39FFC03F80391E000E00 1404B2000E5B120F6C5B6C6C5A3800E0C0013FC7FC191A7E991F>I<3AFFC1FF81FC3A3E 007E0078001E013C13701620001F013E13606C1540A22607804F1380A33A03C0878100A3 3901E103C2A33900F201E4A215F490387C00F8A301381370A301101320261A7F992A> 119 D E /Fe 85 128 df<90381FC1F090387037189038C03E3C3801807C000313783907 003800A9B612C03907003800B2143C397FE1FFC01E2380A21C>11 DII<90380FC07F90397031C0809039E00B0040260180 1E13E00003EB3E013807003C91381C00C01600A7B712E03907001C011500B23A7FF1FFCF FE272380A229>I<1207A2120F121E121C1238126012401280080976A218>19 D<137CEA018738030380000713C0EA0601000E13E0A514C0EB0380A2EB0E00EAFE38EA0E 06EB0380EB01C0A2EB00E014F0A214701478A61470A2EB70E014C0EB718038FE1F001523 7FA218>25 D34 D<7FA2EA03F0EA0C8CEA1082EA208112603840808012C01387A338E0 8300138012701278123F13E0EA1FF8EA07FCEA01FEEA009F1387A2EB8380A2EAF081A312 E00080130012401383EA2086EA1084EA0898EA07E0EA0080A311287DA418>36 DI<133C136213C2EA0181A21203A41382 A213841388EA01C813D09039E001FF809039C0007C006C6C133800011430017013200002 5CEA0478D808385BEA181C003049C7FC38700E02130F38F00704EB038814C8903801F001 EB00E000781370ECB802393C031C06390E0C070C3903F001F021257EA326>I<127012F8 12FCA212741204A41208A21210A212201240060F7CA20E>I<132013401380EA01005A12 061204120CA25AA25AA312701260A312E0AE1260A312701230A37EA27EA2120412067E7E EA0080134013200B327CA413>I<7E12407E7E12187E12041206A27EA2EA0180A313C012 00A313E0AE13C0A312011380A3EA0300A21206A21204120C5A12105A5A5A0B327DA413> I<127012F812FCA212741204A41208A21210A212201240060F7C840E>44 DI<127012F8A3127005057C840E>I<14801301A2EB0300A31306 A35BA35BA35BA35BA35BA3485AA448C7FCA31206A35AA35AA35AA35AA35AA311317DA418 >II<13801203120F12F31203B3 A9EA07C0EAFFFE0F217CA018>III<1303A25BA25B1317A21327136713471387120113071202120612041208A212 101220A2124012C0B512F838000700A7EB0F80EB7FF015217FA018>I<00101380381E07 00EA1FFF5B13F8EA17E00010C7FCA6EA11F8EA120CEA1C07381803801210380001C0A214 E0A4127012F0A200E013C01280EA4003148038200700EA1006EA0C1CEA03F013227EA018 >I<137EEA01C138030080380601C0EA0C03121C381801800038C7FCA212781270A2EAF0 F8EAF30CEAF4067F00F81380EB01C012F014E0A51270A3003813C0A238180380001C1300 EA0C06EA070CEA01F013227EA018>I<12401260387FFFE014C0A23840008038C0010012 801302A2485A5BA25B5BA21360134013C0A21201A25B1203A41207A76CC7FC13237DA118 >III<127012F8A312 701200AB127012F8A3127005157C940E>I<127012F8A312701200AB127012F8A3127812 08A41210A312201240A2051F7C940E>I<497EA3497EA3EB05E0A2EB09F01308A2EB1078 A3497EA3497EA2EBC01F497EA248B51280EB0007A20002EB03C0A348EB01E0A348EB00F0 121C003EEB01F839FF800FFF20237EA225>65 DI<903807E0109038 381830EBE0063901C0017039038000F048C7FC000E1470121E001C1430123CA2007C1410 1278A200F81400A812781510127C123CA2001C1420121E000E14407E6C6C13803901C001 003800E002EB381CEB07E01C247DA223>IIII<903807F00890383C0C 18EBE0023901C001B839038000F848C71278481438121E15185AA2007C14081278A200F8 1400A7EC1FFF0078EB00F81578127C123CA27EA27E7E6C6C13B86C7E3900E0031890383C 0C08903807F00020247DA226>I<39FFFC3FFF390FC003F039078001E0AE90B5FCEB8001 AF390FC003F039FFFC3FFF20227EA125>II<3803FFE038001F007FB3A6127012F8A2130EEAF01EEA401C6C5AEA1870EA 07C013237EA119>IIII<39FF8007FF 3907C000F81570D805E01320EA04F0A21378137C133C7F131F7FEB0780A2EB03C0EB01E0 A2EB00F014F81478143C143E141E140FA2EC07A0EC03E0A21401A21400000E1460121FD8 FFE0132020227EA125>IIIII<3803F020380C0C60EA1802383001E0EA700000601360 12E0A21420A36C1300A21278127FEA3FF0EA1FFE6C7E0003138038003FC0EB07E01301EB 00F0A214707EA46C1360A26C13C07E38C8018038C60700EA81FC14247DA21B>I<007FB5 12F839780780780060141800401408A300C0140C00801404A400001400B3A3497E3801FF FE1E227EA123>I<39FFFC07FF390FC000F86C4813701520B3A5000314407FA200011480 6C7E9038600100EB3006EB1C08EB03F020237EA125>II<3BFFF0 3FFC03FE3B1F8007E000F86C486C48137017206E7ED807801540A24A7E2603C0021480A3 9039E004780100011600A2EC083CD800F01402A2EC101E01785CA2EC200F013C5CA20260 138890391E400790A216D090391F8003F0010F5CA2EC00016D5CA20106130001025C2F23 7FA132>I<397FF803FF390FE001F83907C000E06C6C5B00015CEBF001D800F890C7FCEB 7802EB7C04133EEB1E08EB1F10EB0FB0EB07A014C06D7E130180497EEB0278EB047CEB0C 3EEB081EEB101F496C7E140701407F496C7E1401D801007F486D7E5AD81F807F3AFFC003 FFC022227FA125>II<387FFFFE387E003E0078133C007013781260004013 F012C0EB01E0388003C0A2EB07801200EB0F005B131E5BA25BA25B1201EBE001EA03C0A2 EA07801403EA0F00001E1302A2481306140E48131E00F8137EB512FE18227DA11E>I 92 D97 D<120E12FE121E120EAB131FEB 61C0EB8060380F0030000E1338143C141C141EA7141C143C1438000F1370380C8060EB41 C038083F0017237FA21B>II<14E0130F130113 00ABEA01F8EA0704EA0C02EA1C01EA38001278127012F0A7127012781238EA1801EA0C02 38070CF03801F0FE17237EA21B>II<133E13E33801C780EA0387130748C7FCA9EAFFF80007C7FCB27FEA7FF0112380A2 0F>I<14703803F198380E1E18EA1C0E38380700A200781380A400381300A2EA1C0EEA1E 1CEA33F00020C7FCA212301238EA3FFE381FFFC06C13E0383000F0481330481318A40060 1330A2003813E0380E03803803FE0015217F9518>I<120E12FE121E120EABEB1F80EB60 C0EB80E0380F0070A2120EAF38FFE7FF18237FA21B>I<121C123EA3121CC7FCA8120E12 7E121E120EB1EAFFC00A227FA10E>I<13E0EA01F0A3EA00E01300A81370EA07F0120013 70B3A51260EAF0E013C0EA6180EA3F000C2C83A10F>I<120E12FE121E120EABEB03FCEB 01F014C01480EB02005B5B5B133813F8EA0F1CEA0E1E130E7F1480EB03C0130114E0EB00 F014F838FFE3FE17237FA21A>I<120E12FE121E120EB3ADEAFFE00B237FA20E>I<390E1F C07F3AFE60E183803A1E807201C03A0F003C00E0A2000E1338AF3AFFE3FF8FFE27157F94 2A>I<380E1F8038FE60C0381E80E0380F0070A2120EAF38FFE7FF18157F941B>III<3801F82038070460EA0E02EA1C01003813E0EA7800A25AA71278A2EA 3801121CEA0C02EA070CEA01F0C7FCA9EB0FFE171F7E941A>III<1202A41206A3120E121E123EEAFFFCEA0E00AB1304A6EA07081203EA 01F00E1F7F9E13>I<000E137038FE07F0EA1E00000E1370AD14F0A23806017038038278 3800FC7F18157F941B>I<38FF80FE381E00781430000E1320A26C1340A2EB80C0000313 80A23801C100A2EA00E2A31374A21338A3131017157F941A>I<39FF8FF87F393E01E03C 001CEBC01814E0000E1410EB0260147000071420EB04301438D803841340EB8818141CD8 01C81380EBD00C140E3900F00F00497EA2EB6006EB400220157F9423>I<38FF83FE381F 00F0000E13C06C1380EB8100EA0383EA01C2EA00E41378A21338133C134E138FEA0187EB 0380380201C0000413E0EA0C00383E01F038FF03FE17157F941A>I<38FF80FE381E0078 1430000E1320A26C1340A2EB80C000031380A23801C100A2EA00E2A31374A21338A31310 A25BA35B12F05B12F10043C7FC123C171F7F941A>I<383FFFC038380380EA3007002013 00EA600EEA401C133C1338C65A5B12015B38038040EA07005A000E13C04813805AEA7801 EA7007B5FC12157F9416>III127 D E /Ff 13 118 df<121E123FEA7F80EAFFC0 A4EA7F80EA3F00121E0A0A7A8916>46 D<90391FF8018090B51203000314C73907F007EF 390F8000FF48C7127F003E141F150F5A150712FCA215037EA26C91C7FC13C0EA7FF0EBFF 806C13F8ECFF806C14F06C806C806C14FFC6FC013F1480010114C0D9001F13E01401EC00 3FED1FF0150F1507126000E01403A316E07EA26CEC07C07EB4EC0F8001C0EB1F00D8FBFC 13FE00F1B512F8D8E03F5BD8C003138024317CB02D>83 D<007FB8FCA39039C00FF801D8 7E00EC003F007C82007882A200708200F01780A3481603A5C792C7FCB3AA017FB6FCA331 307DAF38>I87 D97 D100 D104 D108 D<2703F007F8EB1FE000 FFD93FFEEBFFF8913A783F01E0FC02C090388300FE280FF1801FC6137F2607F30013CC01 F602F8148001FC5CA3495CB3B500C3B5380FFFFCA33E207D9F43>I<3903F007F800FFEB 3FFEEC783F02C013803A0FF1801FC03807F30001F614E013FCA35BB3B500C3B5FCA32820 7D9F2D>II<3801FF86000713 FEEA1F00003C133E48131E140E12F8A36C90C7FCB47E13FC387FFFC06C13F0806C7F0007 7F00017FEA003F01001380143F0060131F00E0130FA27E15007E6C131E6C131C38FF8078 38F3FFF038C07F8019207D9F20>115 D 117 D E end %%EndProlog %%BeginSetup %%Feature: *Resolution 300dpi TeXDict begin %%EndSetup %%Page: 1 1 1 0 bop 584 150 a Ff(T)-7 b(uomas)28 b(W.)g(Sandholm)-38 334 y Fe(Departmen)o(t)15 b(of)i(Computer)e(Science)75 394 y(W)l(ashington)j(Univ)o(ersit)o(y)38 455 y(St.)k(Louis,)16 b(MO)g(63130-4899)36 515 y(T)l(elephone:)21 b(\(314\))c(935-4749)105 575 y(F)l(ax:)k(\(314\))c(935-7302)1457 394 y(6951)h(P)o(ershing)e(Av)o (en)o(ue)1398 455 y(Univ)o(ersit)o(y)e(Cit)o(y)l(,)g(MO)33 b(63130)1410 515 y(T)l(elephone:)21 b(\(314\))c(727-2052)577 722 y(In)o(ternet:)j Fd(sandholm@cs.wustl.edu)602 789 y Fe(Born)d(12/21/1968,)i(Helsinki,)14 b(Finland)p -33 867 2017 2 v -38 1116 a Fc(1.)25 b(Educational)18 b(bac)n(kground)p -38 1132 662 2 v -38 1218 a Fe(5/1994{9)q(/19)q(96)86 b Fc(Univ)n(ersit)n(y)17 b(of)i(Massac)n(h)n(usetts)611 b Fe(Amherst,)14 b(MA)362 1278 y(Ph.D.,)h(Computer)g(Science,)g(GP)l(A) h(4.0/4)362 1338 y(Thesis)g(title:)362 1398 y(Negotiation)g(among)g (self-in)o(terested)e(computationally)h(limited)e(agen)o(ts)362 1459 y(Ph.D.)h(Committee:)j(Victor)d(Lesser,)h(James)e(Kurose,)i (Shlomo)f(Zilb)q(erstein)f(\(Com-)362 1519 y(puter)h(Science\);)e(Mark) i(F)l(o)o(x)g(\(Industrial)g(Engineering/Computer)f(Science\);)g(Her-) 362 1579 y(b)q(ert)j(Gin)o(tis)f(\(Economics\))-38 1681 y(9/1992{5)q(/19)q(94)86 b Fc(Univ)n(ersit)n(y)17 b(of)i(Massac)n(h)n (usetts)611 b Fe(Amherst,)14 b(MA)362 1741 y(M.S.,)g(Computer)i (Science,)e(GP)l(A)i(3.95/4)362 1801 y(Master's)f(Pro)s(ject)h(\(2)h (parts\):)362 1861 y(1.)k(Utilit)o(y)14 b(based)j(termination)d(of)j (an)o(ytime)d(algorithms)362 1921 y(2.)21 b(A)16 b(new)g(order)h (parameter)e(for)h(3SA)l(T)-38 2023 y(9/1988{1)q(2/1)q(99)q(1)61 b Fc(Helsinki)16 b(Univ)n(ersit)n(y)h(of)i(T)-5 b(ec)n(hnology)436 b Fe(Esp)q(o)q(o,)18 b(Finland)362 2083 y(M.S.)d(\(B.S.)g(included\))g Fb(with)j(distinction)362 2143 y Fe(Industrial)d(Engineering)h(and)h (Managemen)o(t)e(Science)362 2204 y(Ma)s(jors:)362 2264 y(1.)21 b(Kno)o(wledge)16 b(engineering,)f(GP)l(A)h(4.7/5)362 2324 y(2.)21 b(Business)16 b(strategy)h(and)f(in)o(ternational)g(mark)o (eting,)e(GP)l(A)i(4.2/5)362 2384 y(3.)21 b(Systems)15 b(and)i(op)q(erations)g(researc)o(h,)e(GP)l(A)i(5.0/5)362 2444 y(In)e(addition,)h(18)h(p)q(ost-Master's)h(credits)d(in)h(ma)s (jors)g(1.)21 b(and)c(3.)-38 2546 y(8/1987{7)q(/19)q(88)86 b Fc(Airforce)17 b(Academ)n(y)764 b Fe(Kauha)o(v)m(a,)17 b(Finland)362 2606 y(Pilot)e(Second)i(Lieutenan)o(t)f(\(obligatory)g (military)e(service\))p eop %%Page: 2 2 2 1 bop -38 150 a Fc(2.)25 b(Professional)18 b(exp)r(erience)p -38 166 640 2 v -38 263 a Fe(5/1997{)232 b Fc(BusinessBots,)17 b(Inc.)836 b Fe(P)o(alo)16 b(Alto,)g(CA)362 323 y Fb(Chief)h(Scientist) 362 384 y Fe(BusinessBots,)c(Inc.,)g(a)i(startup)g(compan)o(y)e(based)h (on)h(m)o(ultiagen)o(t)d(tec)o(hnology)l(,)h(is)h(a)362 444 y(mem)n(b)q(er)h(of)j(the)g(CommerceN)o(et)d(consortium)i (consisting)h(of)g(o)o(v)o(er)f(500)i(industrial)362 504 y(partners.)-38 606 y(8/1996{)232 b Fc(W)-5 b(ashington)19 b(Univ)n(ersit)n(y)718 b Fe(St.)21 b(Louis,)c(MO)362 666 y Fb(Assistant)h(Pr)n(ofessor,)e(tenur)n(e)i(tr)n(ack)362 726 y Fe(Departmen)o(t)c(of)j(Computer)e(Science)-38 828 y(9/1992{8)q(/19)q(96)86 b Fc(Univ)n(ersit)n(y)17 b(of)i(Massac)n(h)n(usetts)611 b Fe(Amherst,)14 b(MA)362 888 y Fb(R)n(ese)n(ar)n(ch)i(Assistant)362 948 y Fe(Distributed)f (Problem)g(Solving)h(Lab,)h(Departmen)o(t)d(of)j(Computer)e(Science)-38 1050 y(9/1995{1)q(2/1)q(99)q(5)61 b Fc(Univ)n(ersit)n(y)17 b(of)i(Massac)n(h)n(usetts)611 b Fe(Amherst,)14 b(MA)362 1110 y Fb(T)l(e)n(aching)k(Asso)n(ciate)p Fe(,)e(F)l(ull)f(lecturing)h (resp)q(onsibilit)o(y)362 1170 y(Departmen)o(t)e(of)j(Computer)e (Science)-38 1272 y(12/1990|8/)q(19)q(92)37 b Fc(T)-5 b(ec)n(hnical)18 b(Researc)n(h)g(Cen)n(tre)h(of)g(Finland)343 b Fe(Esp)q(o)q(o,)18 b(Finland)362 1332 y Fb(R)n(ese)n(ar)n(ch)e (Scientist)p Fe(,)h(Lab)q(oratory)i(for)d(Information)f(Pro)q(cessing) -38 1434 y(6/1990{1)q(2/1)q(99)q(0)61 b Fc(Nokia)17 b(Researc)n(h)i (Cen)n(ter)697 b Fe(Esp)q(o)q(o,)18 b(Finland)362 1494 y Fb(R)n(ese)n(ar)n(ch)e(Scientist)p Fe(,)h(Kno)o(wledge)f(Engineering) g(Departmen)o(t)-38 1596 y(12/1988|5/)q(19)q(90)37 b Fc(Kielik)n(one)16 b(Co.)933 b Fe(Esp)q(o)q(o,)18 b(Finland)362 1656 y Fb(Pr)n(o)n(gr)n(ammer)-38 1758 y Fe(6/1989{7)q(/19)q(89)86 b Fc(Smart)17 b(and)i(Final)g(Iris,)e(Co.)614 b Fe(Los)18 b(Angeles,)d(CA)362 1818 y Fb(Datab)n(ase)i(Develop)n(er)-38 1990 y Fc(3.)25 b(Aw)n(ards)20 b(and)f(honors)g(\(gran)n(ts)h(are)e (listed)f(separately)h(later\))p -38 2009 1402 2 v -38 2051 a Fe(\(US$)f(1)f(v)m(aries)h(b)q(et)o(w)o(een)e(4)i(FIM)e(and)i(6) f(FIM\))84 2175 y Fb(NSF)i(CAREER)f(awar)n(d)p Fe(,)e(US$)h(456,000,)i (1997-2002.)84 2277 y Fb(A)o(CM)f(Distinguishe)n(d)h(Dissertation)g(A)o (war)n(d)e(nomination)p Fe(,)h(7/1997.)84 2378 y Fb(Gr)n(aduate)d(F)l (el)r(low)p Fe(,)i(Univ)o(ersit)o(y)10 b(of)j(Massac)o(h)o(usetts,)h (\(a)o(w)o(arded)f(ann)o(ually)f(to)i(one)f(of)g(160)h(grad-)84 2439 y(uate)g(studen)o(ts\))f(US$)g(6,000)h(\(in)f(addition)g(to)g (normal)g(Researc)o(h)f(Assistan)o(t)h(salary\),)g(1994-95.)84 2540 y Fb(De)n(an)21 b(of)g(the)g(Helsinki)h(University)f(of)g(T)l(e)n (chnolo)n(gy)g(A)o(war)n(d)e Fe(for)h(quic)o(k)f(earning)g(of)i(M.S.)d (\(3)84 2600 y(y)o(ears)e(and)h(4)g(mon)o(ths,)e(B.S.)g(included|a)o(v) o(erage)f(is)i(6)h(y)o(ears\),)e(5,000)j(FIM,)d(1991.)p eop %%Page: 3 3 3 2 bop 84 150 a Fe(Conference)13 b(on)h(Inno)o(v)m(ativ)o(e)e (Applications)g(of)i(Arti\014cial)d(In)o(telligence:)17 b(EPO)d(transp)q(ortation)84 210 y(optimization)21 b(system)g(c)o (hosen)h(one)h(of)g(the)f(y)o(ear's)g(17)h(most)f(inno)o(v)m(ativ)o(e)f (AI)h(applications)84 270 y(in)o(ternationally)l(,)15 b(1995.)84 372 y Fb(Finnish)j(Scienc)n(e)i(A)n(c)n(ademy)c(Scholarship) p Fe(,)g(50,000)i(FIM,)d(1993.)84 474 y Fb(Information)j(T)l(e)n (chnolo)n(gy)g(R)n(ese)n(ar)n(ch)e(F)l(oundation)i(Scholarship)p Fe(,)e(25,000)i(FIM,)d(1993.)84 575 y Fb(Finnish)j(Cultur)n(e)g(F)l (oundation)g(R)n(ank)f(Xer)n(ox)h(F)l(und)g(Scholarship)p Fe(,)e(40,000)h(FIM,)e(1995.)84 677 y Fb(Finnish)j(Cultur)n(e)g(F)l (oundation)g(Scholarship)p Fe(,)e(30,000)i(FIM,)d(1991.)84 779 y Fb(T)l(r)n(ansp)n(ortation)i(Ec)n(onomic)h(So)n(ciety)f (Scholarship)p Fe(,)f(15,000)i(FIM,)d(1992.)84 880 y Fb(Exchange)26 b(R)n(ese)n(ar)n(cher)c(Scholarship)p Fe(,)i(T)l(ec)o(hnical)e(Researc)o(h)g(Cen)o(tre)g(of)h(Finland,)h (186,000)84 941 y(FIM,)15 b(1992.)84 1042 y Fb(Heikki)k(and)f(Hilma)g (Honkanen)h(F)l(oundation)f(Me)n(dal)g(and)f(Scholarship)p Fe(,)f(30,000)i(FIM,)d(1994.)84 1144 y Fb(Charles)j(Baudoin)g (Scholarship)p Fe(,)e(Thanks)h(to)g(Scandina)o(via)e(F)l(oundation,)i (US$)f(3,000,)h(1991.)84 1246 y Fb(L)n(e)n(o)g(and)g(R)n(e)n(gina)g (Wainstein)i(F)l(oundation)f(Scholarship)p Fe(,)e(40,000)i(FIM,)d (1994.)84 1347 y Fb(Jenny)j(and)g(A)o(ntti)g(Wihuri)f(F)l(oundation)h (Scholarship)p Fe(,)e(30,000)i(FIM,)d(1993.)84 1449 y Fb(El)r(la)k(and)f(Ge)n(or)n(ge)f(Ehrnr)n(o)n(oth)f(F)l(oundation)i (Scholarship)p Fe(,)e(25,000)i(FIM,)c(1994.)84 1551 y Fb(T)l(r)n(avel)i(Scholarship)p Fe(,)e(In)o(ternational)f(Join)o(t)g (Conference)g(on)h(Arti\014cial)e(In)o(telligence)f(\(IJCAI\),)84 1611 y(US$)17 b(800,)g(1997.)84 1713 y Fb(Student)f(A)o(uthor)d (Scholarship)p Fe(,)f(Second)h(In)o(ternational)e(Conference)g(on)i (Multiagen)o(t)e(Systems)84 1773 y(\(ICMAS\),)k(US$)i(1000,)g(1996)h (\(declined)c(b)q(ecause)j(not)g(a)f(studen)o(t)g(an)o(ymore\).)84 1875 y Fb(Student)f(A)o(uthor)d(Scholarship)p Fe(,)f(13th)h(National)f (Conference)f(on)h(Arti\014cial)e(In)o(telligence)f(\(AAAI\),)84 1935 y(US$)17 b(610,)g(1996.)84 2036 y Fb(Student)j(A)o(uthor)d (Scholarship)p Fe(,)f(14th)h(In)o(ternational)f(Join)o(t)g(Conference)g (on)h(Arti\014cial)e(In)o(tel-)84 2097 y(ligence)g(\(IJCAI\),)g(US$)h (500,)h(1995.)84 2198 y Fb(Student)k(A)o(uthor)e(Scholarship)p Fe(,)f(First)g(In)o(ternational)f(Conference)g(on)i(Multiagen)o(t)e (Systems)84 2258 y(\(ICMAS\),)e(US$)i(600,)g(1995.)84 2360 y Fb(Student)28 b(A)o(uthor)d(Scholarship)p Fe(,)j(AAAI)c(Spring)h (Symp)q(osium:)38 b(Arti\014cial)23 b(In)o(telligence)g(in)84 2420 y(Medicine,)14 b(US$)j(600,)g(1996.)84 2522 y Fb(Student)23 b(A)o(uthor)e(Scholarship)p Fe(,)h(Engineering)e(F)l(oundation,)h(In)o (ternational)f(Conference)g(on)84 2582 y(Arti\014cial)15 b(In)o(telligence)e(Applications)j(in)g(T)l(ransp)q(ortation)i (Engineering,)d(US$)i(600,)g(1992.)p eop %%Page: 4 4 4 3 bop 84 150 a Fb(Pr)n(o)22 b(Scientia)j(Me)n(dal)p Fe(,)e(Asso)q(ciation)f(for)h(T)l(eac)o(hers)e(of)i(Mathematical)d (Studies,)j(600)g(FIM,)84 210 y(1987.)84 312 y Fb(Writing)18 b(c)n(omp)n(etition)f(awar)n(d)p Fe(,)e(League)i(of)g(Finnish-American) d(So)q(cieties,)g(1987.)84 413 y Fb(Chemistry)j(Olympics)h(tr)n(aine)n (e)p Fe(,)e(1986.)-38 586 y Fc(4.)25 b(Professional)18 b(activities)p -38 593 601 2 v -38 693 a Fe(T)l(utorials)84 802 y Fb(A)n(gents)26 b(in)f(Ele)n(ctr)n(onic)g(Markets)p Fe(.)45 b(8.5)24 b(hour)g(funded)g(tutorial)g(at)h(the)e(Horizon)h (Systems)84 862 y(Lab)q(oratory)l(,)18 b(Mitsubishi)d(Electric)f (Information)h(T)l(ec)o(hnology)h(Cen)o(ter)f(America,)e(W)l(altham,)84 923 y(MA,)j(10/15/1997.)84 1024 y Fb(Ec)n(onomic)n(al)r(ly)22 b(F)l(ounde)n(d)g(Multiagent)h(Systems)p Fe(.)34 b(4)20 b(hour)h(funded)f(tutorial)h(to)f(b)q(e)h(giv)o(en)e(at)84 1085 y(the)13 b(In)o(ternational)g(Join)o(t)f(Conference)h(on)g (Arti\014cial)f(In)o(telligence)e(\(IJCAI\),)i(Nago)o(y)o(a,)h(Japan,) 84 1145 y(8/23/199)q(7.)84 1246 y Fb(A)n(gents)h(in)f(Ele)n(ctr)n(onic) g(Markets)p Fe(.)20 b(8.5)11 b(hour)h(funded)f(tutorial)g(at)g(the)g (Agen)o(ts)f(Group,)j(Hewlett-)84 1307 y(P)o(ac)o(k)m(ard)k(Labs,)g (Bristol,)e(UK,)g(8/5/1997-8)q(/6)q(/19)q(97.)84 1408 y Fb(A)n(gents)20 b(in)g(Ele)n(ctr)n(onic)f(Markets)p Fe(.)26 b(1.5)18 b(hour)h(funded)f(tutorial)f(at)h(BusinessBots,)g (Inc,)f(\(Mit-)84 1468 y(subishi)g(Electric)d(premises,)g(San)j(Jose,)f (CA\),)g(5/19/1997.)84 1570 y Fb(A)n(gents)23 b(in)f(Ele)n(ctr)n(onic)h (Markets)p Fe(.)35 b(8.5)21 b(hour)g(funded)g(tutorial)g(at)g(the)g (First)f(In)o(ternational)84 1630 y(Conference)c(on)h(Autonomous)f (Agen)o(ts)g(\(Agen)o(ts-97\),)g(Marina)g(del)g(Rey)l(,)f(CA,)h (2/5/1997.)-38 1732 y(Program)h(committe)o(e)c(mem)o(b)q(er)84 1841 y(National)k(Conference)e(on)i(Arti\014cial)e(In)o(telligence)e (\(AAAI\),)h(Madison,)i(WI,)g(1998)84 1943 y(National)h(Conference)e (on)i(Arti\014cial)e(In)o(telligence)e(\(AAAI\),)h(Pro)o(vidence,)g (RI,)i(1997)84 2044 y(National)h(Conference)e(on)i(Arti\014cial)e(In)o (telligence)e(\(AAAI\),)h(P)o(ortland,)i(OR,)g(1996)84 2146 y(NSF)g(electronic)f(commerc)o(e)e(initiativ)o(e)h(planning)j (committe)o(e,)c(1997.)84 2248 y(18th)20 b(In)o(ternational)f (Conference)f(on)i(Distributed)f(Computing)f(Systems)g(\(ICDCS\),)h (Ams-)84 2308 y(terdam,)c(The)h(Netherlands,)f(1998)84 2410 y(Third)22 b(In)o(ternational)f(Conference)g(on)h(Multi-Agen)o(t)e (Systems)g(\(ICMAS\),)g(P)o(aris,)i(F)l(rance,)84 2470 y(1998)84 2572 y(Second)11 b(In)o(ternational)g(Conference)f(on)h (Autonomous)g(Agen)o(ts)g(\(Agen)o(ts\),)g(Minneap)q(olis,)g(MN,)84 2632 y(1998)p eop %%Page: 5 5 5 4 bop 84 150 a Fe(First)21 b(In)o(ternational)g(Conference)g(on)h (Autonomous)f(Agen)o(ts)g(\(Agen)o(ts\),)h(Marina)f(del)g(Rey)l(,)84 210 y(CA,)16 b(1997)84 312 y(Second)24 b(In)o(ternational)f(Conference) g(on)h(Autonomous)g(Agen)o(ts)f(\(Agen)o(ts\),)h(W)l(orkshop)h(on)84 372 y(Agen)o(t-Mediated)15 b(Electronic)g(T)l(rading,)i(Minneap)q (olis,)e(MN,)g(1998)84 474 y(Second)24 b(In)o(ternational)f(Conference) g(on)h(Autonomous)g(Agen)o(ts)f(\(Agen)o(ts\),)h(W)l(orkshop)h(on)84 534 y(Arti\014cial)15 b(So)q(cieties)g(and)i(Computational)f(Mark)o (ets,)f(Minneap)q(olis,)h(MN,)f(1998)84 635 y(AAAI)k(Spring)i(Symp)q (osium:)27 b(Satis\014cing)21 b(Mo)q(dels,)g(Stanford,)h(CA,)e(1998.)35 b(\(One)20 b(of)h(eigh)o(t)84 696 y(organizers\))84 797 y(In)o(ternational)e(W)l(orkshop)i(on)f(Co)q(op)q(erativ)o(e)g (Information)e(Agen)o(ts)h(\(CIA\):)g(Learning,)h(Mo-)84 857 y(bilit)o(y)l(,)i(and)h(Electronic)f(Commerce)d(for)k(Information)f (Searc)o(h)g(in)g(the)h(Infosphere,)g(P)o(aris,)84 918 y(F)l(rance,)16 b(1998)84 1019 y(In)o(ternational)11 b(W)l(orkshop)g(on)h(Co)q(op)q(erativ)o(e)f(Information)f(Agen)o(ts)g (\(CIA\):)g(D)o(AI)g(Meets)g(Database)84 1080 y(Systems,)15 b(Kiel,)f(German)o(y)l(,)h(1997)84 1181 y(Scandina)o(vian)i(Conference) e(on)i(Arti\014cial)d(In)o(telligence)g(\(SCAI\),)h(Helsinki,)f (Finland,)h(1997)-38 1283 y(Chair)84 1392 y(Co-c)o(hair)h(of)g(the)f Fb(Market-b)n(ase)n(d)i(Systems)f Fe(w)o(orking)g(group)g(at)g(the)g (In)o(ternational)f(W)l(orkshop)84 1452 y(on)i(Multi-Agen)o(t)e (Systems)g(\(IWMAS\),)g(MIT,)g(Octob)q(er)h(12-14,)i(1997.)84 1554 y(Discussion)h(leader)e(of)h(the)g Fb(Using)i(Ec)n(onomics)f(for)f (R)n(esour)n(c)n(e)g(A)o(l)r(lo)n(c)n(ation)i(and)f(Contr)n(ol)f Fe(ses-)84 1614 y(sion.)i(Second)11 b(In)o(ternational)f(W)l(orkshop)i (on)f(Mobile)f(Agen)o(ts.)19 b(Dartmouth)11 b(College,)g(Hano)o(v)o (er,)84 1674 y(NH,)k(Septem)o(b)q(er)g(19-20,)i(1997)84 1776 y(In)o(ternational)i(Join)o(t)g(Conference)g(on)h(Arti\014cial)e (In)o(telligence)f(\(IJCAI\),)h(Session)h(K3:)28 b(Dis-)84 1836 y(tributed)16 b(AI,)f(Nago)o(y)o(a,)h(Japan,)h(8/29/1997)84 1938 y(First)11 b(In)o(ternational)f(Conference)g(on)i(Autonomous)e (Agen)o(ts)h(\(Agen)o(ts-97\),)h(Session)f(2:)19 b(Kno)o(wl-)84 1998 y(edge)e(Acquisition)d(and)j(Accum)o(ulation,)c(Marina)k(del)e (Rey)l(,)g(CA,)h(1997)84 2100 y(National)h(Conference)g(on)h (Arti\014cial)d(In)o(telligence)f(\(AAAI\),)h(Session)j(on)f(Constrain) o(t)h(Satis-)84 2160 y(faction:)k(Sto)q(c)o(hastic)16 b(Searc)o(h,)f(P)o(ortland,)h(OR,)g(1996)84 2262 y(AI)j(and)h (Logistics)g(W)l(orkshop,)h(Finnish)e(Arti\014cial)f(In)o(telligence)f (Conference,)i(\(Organizer,)84 2322 y(Chair,)d(and)h(Referee\),)e (Helsinki,)e(Finland,)j(1992)-38 2423 y(Review)g(panelist)84 2532 y(NSF)k(ITO)g(\(Information)f(T)l(ec)o(hnology)h(and)g (Organizations\))h(gran)o(t)f(review)f(panel,)i(W)l(ash-)84 2593 y(ington,)c(DC)f(\(7)h(reviews\))p eop %%Page: 6 6 6 5 bop -38 150 a Fe(In)o(vited)15 b(participan)o(t)84 259 y(Institute)g(for)g(Op)q(erations)g(Researc)o(h)g(and)g(the)g (Managemen)o(t)f(Sciences)f(\(INF)o(ORMS\))g(In)o(ter-)84 319 y(national)k(conference.)j(Mon)o(treal,)15 b(Canada,)j(April)d (26-29,)i(1998.)84 421 y(In)o(ternational)f(W)l(orkshop)h(on)g (Multiagen)o(t)e(Systems.)20 b(MIT,)15 b(Octob)q(er)i(12-14,)g(1997.)84 522 y(Second)22 b(In)o(ternational)f(W)l(orkshop)i(on)f(Mobile)e(Agen)o (ts.)37 b(Dartmouth)22 b(College,)g(Hano)o(v)o(er,)84 582 y(NH,)15 b(Septem)o(b)q(er)g(19-20,)i(1997)h(\(funded)e(participan) o(t\).)84 684 y(D)o(ARP)l(A)i(Y)l(oung)h(In)o(v)o(estigator)f(W)l (orkshop,)h(Pro)o(vidence,)f(RI,)f(July)h(27-28,)j(1997.)29 b(\(The)19 b(20)84 744 y(researc)o(hers)d(w)o(ere)f(selected)f(from)h (o)o(v)o(er)g(100)i(p)q(eople)f(nominated)f(b)o(y)h(senior)g(mem)n(b)q (ers)e(of)i(the)84 805 y(AI)g(comm)o(unit)n(y\).)84 906 y(Decen)o(tralization)j(Conference)g(\(NSF)g(sp)q(onsored\),)j(P)o(enn) d(State)h(Univ)o(ersit)o(y)l(,)d(Departmen)o(t)84 966 y(of)g(Economics,)e(Ma)o(y)g(9-11,)j(1997)f(\(funded)g(participan)o (t\).)-38 1068 y(Referee)84 1177 y(National)g(Science)d(F)l(oundation)j (\(NSF\),)f(1997)84 1279 y(Arti\014cial)f(In)o(telligence)e (\(Journal\),)k(1995,)g(1996,)g(1997)84 1380 y(IEEE)g(T)l(ransactions)g (on)g(P)o(attern)f(Analysis)g(and)h(Mac)o(hine)e(In)o(telligence)e (\(Journal\),)j(1997)84 1482 y(Journal)h(of)g(Arti\014cial)d(In)o (telligence)g(Researc)o(h,)h(1996)j(\(2\),)e(1997)i(\(3\))84 1584 y(IEEE)f(T)l(ransactions)g(on)g(Systems,)d(Man,)i(and)h(Cyb)q (ernetics)f(\(Journal\),)g(1997)84 1686 y(Computational)g(In)o (telligence)e(\(Journal\),)i(1996)84 1787 y(IEEE)h(Exp)q(ert)f (\(Journal\),)g(1994)84 1889 y(Arti\014cial)j(In)o(telligence)e (\(Journal\))j(Sp)q(ecial)f(Issue)h(on)g(Principles)e(of)j(Multiagen)o (t)d(Systems,)84 1949 y(1995)84 2051 y(IEEE)f(T)l(ransactions)g(on)g (Data)g(and)g(Kno)o(wledge)f(Engineering)g(\(Journal\),)g(1996)84 2153 y(Journal)h(of)f(Exp)q(erimen)o(tal)d(and)j(Theoretical)f (Arti\014cial)f(In)o(telligence)f(\(JET)l(AI\),)i(1996,)i(1997)84 2254 y(In)o(ternational)f(Join)o(t)g(Conference)g(on)g(Arti\014cial)f (In)o(telligence)e(\(IJCAI\),)i(1995,)j(1997)84 2356 y(National)f(Conference)e(on)i(Arti\014cial)e(In)o(telligence)e (\(AAAI\),)h(1996,)j(1997,)h(1998)84 2458 y(In)o(ternational)e (Conference)g(on)g(Mac)o(hine)g(Learning)g(\(ICML\),)g(1996)84 2559 y(IEEE)h(INF)o(OCOM,)d(The)i(Conference)g(on)h(Computer)e(Comm)o (unications,)e(1998)p eop %%Page: 7 7 7 6 bop 84 150 a Fe(First)13 b(In)o(ternational)g(Conference)g(on)h (Multiagen)o(t)f(Systems)f(\(ICMAS\),)g(San)i(F)l(rancisco,)f(CA,)84 210 y(1995)84 312 y(First)h(In)o(ternational)g(Conference)f(on)i (Autonomous)f(Agen)o(ts)g(\(Agen)o(ts-97\),)h(Marina)f(del)g(Rey)l(,)84 372 y(CA,)i(1997)84 474 y(Scandina)o(vian)e(Conference)f(on)i (Arti\014cial)d(In)o(telligence)f(\(SCAI\),)i(Sp)q(ecial)g(theme)f(of)i (the)g(con-)84 534 y(ference)h(w)o(as)i(In)o(telligen)o(t)d(Agen)o(ts,) h(Helsinki,)f(Finland,)h(1997)84 635 y(In)o(ternational)22 b(Join)o(t)f(Conference)g(on)i(Arti\014cial)d(In)o(telligence)f(W)l (orkshop)k(on)f(Adaptation)84 696 y(and)17 b(Learning)g(in)f(Multiagen) o(t)f(Systems,)g(Mon)o(treal,)g(Canada,)i(1995)84 797 y(Europ)q(ean)k(Conference)d(on)h(Arti\014cial)f(In)o(telligence)e(W)l (orkshop)k(on)f(Learning)h(in)e(D)o(AI)g(Sys-)84 857 y(tems,)d(Budap)q(est,)h(Hungary)l(,)g(1996)84 959 y(AAAI)f(Spring)i (Symp)q(osium)e(Series:)21 b(Adaptation,)c(Co)q(ev)o(olution)f(and)i (Learning)f(in)f(Multia-)84 1019 y(gen)o(t)g(Systems,)f(Stanford)i (Univ)o(ersit)o(y)l(,)c(CA,)j(1996)-38 1121 y(App)q(earances)h(in)f (the)g(national)h(media)84 1230 y(In)o(terview)o(ed)f(for)i(and)h (cited)e(in)g(the)h Fb(New)i(Y)l(ork)g(Times)e Fe(2/10/1997:)28 b(\\In)o(telligen)o(t)16 b(Soft)o(w)o(are)84 1290 y(Finding)g(Nic)o (he".)84 1392 y(In)o(terview)o(ed)21 b(for)i(and)g(sp)q(ok)o(e)h(on)f (the)f Fb(Scienc)n(e)k(Up)n(date)d Fe(US)f(national)i(radio)f(program)g (re-)84 1452 y(garding)d(electronic)d(commerce)e(systems)j(consisting)h (of)g(m)o(ultiple)d(self-in)o(terested)h(soft)o(w)o(are)84 1512 y(agen)o(ts.)k(O\016cial)13 b(broadcast)i(date)g(w)o(as)f (7/18/97,)j(but)d(actual)g(broadcast)h(dates)g(v)m(ary)f(station)84 1572 y(to)j(station.)22 b(Also)16 b(broadcast)h(on)g(Mutual)f (Broadcasting's)h Fb(A)o(meric)n(a)g(in)h(the)g(Morning)p Fe(.)84 1674 y Fb(Who'sWho)f(in)g(Scienc)n(e)j(and)d(Engine)n(ering)i Fe(US)c(\(Marquis\),)h(4th)h(edition.)-38 1776 y(Other)f(education)h (related)e(service)84 1885 y(Mem)o(b)q(er,)f(Helsinki)g(Univ)o(ersit)o (y)g(of)i(T)l(ec)o(hnology)g(Studen)o(t)g(Bo)q(dy)h(P)o(arliamen)o(t)84 1986 y(Revisor,)f(Finnish)f(Arti\014cial)g(In)o(telligence)e(So)q(ciet) o(y)84 2088 y(Revisor,)j(Pro)q(duction)h(Econom)o(y)e(Guild)84 2190 y(F)l(ounding)i(Mem)o(b)q(er,)d(Europ)q(ean)j(Studen)o(ts)f(of)h (Industrial)f(Engineering)g(and)h(Managemen)o(t)84 2292 y(Mem)o(b)q(er,)d(AIESEC)i(HUT)g(Executiv)o(e)e(Committee)84 2393 y(Captain,)j(Pro)q(duction)g(Econom)o(y)e(Guild)h(F)l(reshman)f (Activities)84 2495 y(Mem)o(b)q(er,)f(Mac)o(hine)h(Engineer)h(Guild)g (Go)o(v)o(ernmen)o(t)84 2597 y(Mem)o(b)q(er,)e(Kulosaari)j (Co-Educational)g(Sc)o(ho)q(ol)g(Board)p eop %%Page: 8 8 8 7 bop 84 150 a Fe(Chairman,)16 b(Kulosaari)g(Co-Educational)i(Sc)o (ho)q(ol,)e(High)g(Sc)o(ho)q(ol)g(Directariat)-38 251 y(Professional)h(a\016liations)84 360 y(American)d(Asso)q(ciation)j (for)f(Arti\014cial)f(In)o(telligence)84 462 y(Asso)q(ciation)i(for)g (Computing)e(Mac)o(hinery)84 564 y(Europ)q(ean)j(So)q(ciet)o(y)d(for)i (Arti\014cial)d(Neural)i(Net)o(w)o(orks)84 665 y(Finnish)g (Arti\014cial)f(In)o(telligence)e(Researc)o(h)j(So)q(ciet)o(y)-38 838 y Fc(5.)25 b(Courses)19 b(taugh)n(t)p -38 854 440 2 v -38 934 a Fe(Marc)o(h)d(1998)152 b Fc(Multiagen)n(t)18 b(Systems)573 b Fe(Bahia)17 b(Blanca,)e(Argen)o(tina)362 994 y Fb(Universidad)i(Nacional)i(del)f(Sur)362 1054 y Fe(In)o(vited)c(funded)i(in)o(tensiv)o(e)f(course)h(for)g(15)h(da)o (ys)-38 1153 y(Spring)g(1998)149 b Fc(CS)19 b(511:)24 b(Arti\014cial)18 b(In)n(telligence)e(I)512 b Fe(St.)21 b(Louis,)c(MO)362 1213 y Fb(Assistant)h(Pr)n(ofessor,)e(Washington)i (University)-38 1312 y Fe(Spring)f(1998)149 b Fc(CS)18 b(6745:)24 b(Researc)n(h)17 b(Seminar)f(on)i(Arti\014cial)e(In)n(tel-) 362 1372 y(ligence)1643 1312 y Fe(St.)21 b(Louis,)c(MO)362 1421 y Fb(Assistant)h(Pr)n(ofessor,)e(Washington)i(University)-38 1520 y Fe(F)l(all)e(1997)208 b Fc(CS)19 b(520A:)f(In)n(telligen)n(t)g (Real-Time)d(Systems)287 b Fe(St.)21 b(Louis,)c(MO)362 1580 y Fb(Assistant)h(Pr)n(ofessor,)e(Washington)i(University)362 1640 y Fe(Generated)e(new)g(course)362 1700 y(T)l(opics)24 b(included:)36 b(mo)q(dels)23 b(for)h(represen)o(ting)f(computational)h (limitations)e(and)362 1760 y(tradeo\013s,)27 b(decision)c(theory)i (and)g(rational)f(c)o(hoice,)h(the)f(v)m(alue)g(of)h(information,)362 1821 y(the)f(delib)q(erativ)o(e)e(vs.)46 b(reactiv)o(e)23 b(debate,)j(principles)d(of)i(meta-reasoning,)h(real-)362 1881 y(time)17 b(searc)o(h,)i(memory-b)q(ounded)e(searc)o(h,)j(utilit)o (y-directed)c(searc)o(h,)j(delib)q(eration)362 1941 y(sc)o(heduling)9 b(\(con)o(trol)i(of)g(reasoning\),)h(soft)g(real-time,)d(an)o(ytime)g (algorithms,)h(design-)362 2001 y(to-time)16 b(algorithms,)h(dynamic)f (planning)i(and)h(execution,)e(reinforcemen)o(t)e(learn-)362 2061 y(ing,)g(and)i(ev)m(aluation)g(of)f(resource-b)q(ounded)h (reasoning)g(tec)o(hniques)-38 2160 y(F)l(all)f(1997)208 b Fc(CS)18 b(6745:)24 b(Researc)n(h)17 b(Seminar)f(on)i(Arti\014cial)e (In)n(tel-)362 2220 y(ligence)1643 2160 y Fe(St.)21 b(Louis,)c(MO)362 2269 y Fb(Assistant)h(Pr)n(ofessor,)e(Washington)i(University)-38 2368 y Fe(Spring)f(1997)149 b Fc(CS)18 b(6745:)24 b(Researc)n(h)17 b(Seminar)f(on)i(Arti\014cial)e(In)n(tel-)362 2428 y(ligence)1643 2368 y Fe(St.)21 b(Louis,)c(MO)362 2476 y Fb(Assistant)h(Pr)n(ofessor,) e(Washington)i(University)362 2537 y Fe(T)l(opics)23 b(included:)34 b(Reasoning)24 b(ab)q(out)g(kno)o(wledge,)g(recursiv)o (e)e(agen)o(t)h(mo)q(deling,)362 2597 y(resource-b)q(ounded)17 b(reasoning,)f(game)g(theory)l(,)f(and)i(coalition)f(formation)p eop %%Page: 9 9 9 8 bop -38 150 a Fe(F)l(all)16 b(1996)208 b Fc(CS)19 b(516A:)f(Multiagen)n(t)h(Systems)541 b Fe(St.)21 b(Louis,)c(MO)362 210 y Fb(Assistant)h(Pr)n(ofessor,)e(Washington)i(University)362 270 y Fe(Generated)e(new)g(course)362 330 y(Collected)i(readings)i (from)f(a)h(v)m(ariet)o(y)f(of)h(sources)h(and)f(wrote)g(handouts)h(b)q (ecause)362 390 y(no)16 b(adequate)h(textb)q(o)q(ok)f(exists)362 451 y(Constructed)g(and)h(graded)g(three)f(homew)o(ork)f(sets)h (\(written)g(and)g(programming\))362 511 y(Co-writing)g(a)h(textb)q(o)q (ok)g(for)f(this)h(class)f(with)g(in)o(ternational)g(colleagues)-38 610 y(F)l(all)g(1995)208 b Fc(CS)19 b(187:)24 b(Programming)16 b(with)j(Data)g(Structures)205 b Fe(Amherst,)14 b(MA)362 670 y Fb(T)l(e)n(aching)k(Asso)n(ciate,)g(University)g(of)f (Massachusetts)362 730 y Fe(Ga)o(v)o(e)e(all)h(lectures)f(\(3/w)o(eek,) h(50)h(min.)i(eac)o(h\))d(to)g(40)h(studen)o(t)f(section)362 790 y(Generated)g(exam)e(questions)362 851 y(Constructed)24 b(solutions)g(and)g(graded)h(programming)d(and)i(written)f(homew)o (orks)362 911 y(and)16 b(exams)362 971 y(Professor)h(in)f(c)o(harge:)21 b(P)o(aul)16 b(Utgo\013)-38 1070 y Fc(Other)j(T)-5 b(eac)n(hing)19 b(Service)-38 1190 y Fe(1986)304 b Fb(Computer)17 b(English)j Fc(education)e(video)437 b Fe(Helsinki,)14 b(Finland)362 1250 y Fb(Particip)n(ant)362 1311 y Fe(Used)h(in)h(Finland)g(to)h(teac) o(h)e(computer)g(literacy)g(in)h(English)g(at)h(B.S.)e(lev)o(el)-38 1410 y Fc(Education)k(in)f(T)-5 b(eac)n(hing)-38 1530 y Fe(1995)304 b Fc(Univ)n(ersit)n(y)17 b(of)i(Massac)n(h)n(usetts)611 b Fe(Amherst,)14 b(MA)362 1590 y Fb(T)l(e)n(acher)j(orientation)h(day) -38 1689 y Fe(1991)304 b Fc(Helsinki)16 b(Univ)n(ersit)n(y)h(of)i(T)-5 b(ec)n(hnology)436 b Fe(Esp)q(o)q(o,)18 b(Finland)362 1749 y(Semester)c(long)i Fb(Conducting)j(Smal)r(l)g(Gr)n(oup)e(A)n (ctivities)g Fe(course)362 1809 y(\(practicum\))-38 1908 y(1989)304 b Fc(Helsinki)16 b(Univ)n(ersit)n(y)h(of)i(T)-5 b(ec)n(hnology)436 b Fe(Esp)q(o)q(o,)18 b(Finland)362 1969 y(Semester)c(long)i Fb(Sp)n(e)n(e)n(ch)i(Communic)n(ation)e Fe(course)-38 2138 y Fc(6.)25 b(Sup)r(ervising)p -38 2155 354 2 v 35 2276 a Fa(\017)f Fe(Martin)16 b(Andersson)h(\(PhD)g (studen)o(t\),)e(10/96{)35 2377 y Fa(\017)24 b Fe(Kate)17 b(Larsson)h(\(PhD)e(studen)o(t\),)g(8/97{)35 2479 y Fa(\017)24 b Fe(Qian)o(b)q(o)17 b(Huai)f(\(PhD)h(studen)o(t\),)e(8/97{)35 2581 y Fa(\017)24 b Fe(Dr.)45 b(F)l(ernando)25 b(T)l(ohm)o(\023)-23 b(e)23 b(\(researc)o(h)g(asso)q(ciate,)j(p)q(ostdo)q(ctoral,)i(on)c (lea)o(v)o(e)e(from)h(his)h(p)q(osition)h(as)84 2641 y(ten)o(ured)15 b(Asso)q(ciate)h(Professor)h(of)f(Mathematical)e (Economics)g(in)i(Univ)o(ersidad)e(Nacional)i(del)f(Sur,)p eop %%Page: 10 10 10 9 bop 84 150 a Fe(Bahia)17 b(Blanca,)e(Argen)o(tina\),)g(8/96{)35 251 y Fa(\017)24 b Fe(Dr.)e(Sviatosla)o(v)16 b(Braino)o(v)f(\(researc)o (h)h(asso)q(ciate,)h(p)q(ostdo)q(ctoral\),)g(1/98{)-38 424 y Fc(7.)25 b(Departmen)n(t/Sc)n(ho)r(ol/Comm)m(unit)n(y)16 b(service)p -38 443 1052 2 v -38 543 a Fe(1997)304 b Fc(W)-5 b(ashU)19 b(CS)g(MS)g(pro)s(ject)f(committe)o(e)e(of)j(Marin)f (Bezic)160 b Fe(St.)21 b(Louis,)c(MO)362 603 y Fb(Memb)n(er)362 664 y Fe(Title:)22 b(Design)c(and)g(Implem)o(en)n(tation)d(of)j(a)g (Distributed)e(Data)j(Exploration)f(and)g(Pro-)362 724 y(cessing)e(System)362 784 y(Graduated)h(F)l(all-97)-38 886 y(7/1997{)232 b Fc(W)-5 b(ashU)19 b(CS)g(Graduate)g(Studen)n(t)g (Admissions)e(Committee)64 b Fe(St.)21 b(Louis,)c(MO)362 946 y Fb(Chair)-38 1048 y Fe(7/1997{)232 b Fc(W)-5 b(ashU)19 b(In)n(ternational)f(Exc)n(hange)h(Committe)o(e)322 b Fe(St.)21 b(Louis,)c(MO)362 1108 y Fb(R)n(epr)n(esentative)h(of)f(the)h (Dep)n(artment)f(of)h(Computer)f(Scienc)n(e)-38 1209 y Fe(7/1997{)232 b Fc(W)-5 b(ashU)19 b(CS)g(Master's)f(Program)g (Committee)358 b Fe(St.)21 b(Louis,)c(MO)362 1270 y Fb(Memb)n(er)-38 1371 y Fe(6/1997{7)q(/19)q(97)86 b Fc(W)-5 b(ashU)17 b(CS)h(PhD)f(Candidacy)g(Quali\014er)f(System)f(Revi-)362 1432 y(sion)j(Ad)h(Ho)r(c)f(Committee)1743 1371 y Fe(St.)j(Louis,)c(MO) 362 1470 y Fb(One)h(of)f(thr)n(e)n(e)g(memb)n(ers)p Fe(.)k(New)16 b(system)f(instan)o(tiated)-38 1572 y(8/1996{)232 b Fc(W)-5 b(ashU)19 b(CS)g(Graduate)g(Studen)n(t)g(Admissions)e(Committee)64 b Fe(St.)21 b(Louis,)c(MO)362 1632 y Fb(Memb)n(er)-38 1734 y Fe(1/28/199)q(7)183 b Fc(W)-5 b(ashU)19 b(CS)g(Oral)g (Quali\014er,)e(Ram)h(Seth)n(uraman)294 b Fe(St.)21 b(Louis,)c(MO)362 1794 y Fb(One)h(of)f(thr)n(e)n(e)g(memb)n(ers)-38 1896 y Fe(8/1996{)232 b Fc(W)-5 b(ashU)19 b(CS)g(AI)g(curriculum)e (enhancemen)n(t)g(and)362 1956 y(mo)r(dernization)e(planning)1743 1896 y Fe(St.)21 b(Louis,)c(MO)362 2004 y Fb(A)o(lone)h(and)g(with)g (Asso)n(c.)k(Pr)n(of.)f(L)n(oui)c(and)g(Asso)n(c.)23 b(Pr)n(of.)e(Gil)r(lett)362 2065 y(T)l(otaling)d(ab)n(out)g(25)f(hours) g(by)g(1/12/97.)-38 2166 y Fe(8/19/199)q(6)183 b Fc(Presen)n(tation)18 b(at)g(the)h(facult)n(y)f(retreat)535 b Fe(St.)21 b(Louis,)c(MO)362 2227 y Fb(\\Up)n(dating)g(&)h(enlar)n(ging)h(the)f(AI)f(pr)n(o)n(gr)n (am)e(at)j(WashU)f(CS")-38 2328 y Fe(10/1996)232 b Fc(Sc)n(ho)r(ol)18 b(District)f(of)i(Univ)n(ersit)n(y)e(Cit)n(y)362 2388 y(Gifted)g(and)j(T)-5 b(alen)n(ted)18 b(Education)g(Program)1616 2328 y Fe(Univ)o(ersit)o(y)c(Cit)o(y)l(,)h(MO)362 2437 y Fb(One)j(hour)f(talk)h(to)g(5th)f(gr)n(ade)h(students)g(on)g(their)f (international)i(day)p eop %%Page: 11 11 11 10 bop -38 150 a Fc(8.)25 b(Consulting)19 b(w)n(ork)p -38 166 472 2 v -38 287 a Fe(5/1997{)232 b Fc(BusinessBots,)17 b(Inc.)836 b Fe(P)o(alo)16 b(Alto,)g(CA)362 347 y Fb(Chief)h(Scientist) 362 407 y Fe(P)o(aten)o(ting)11 b(and)i(\014elding)f(m)o(ultiagen)o(t)d (negotiation)k(tec)o(hniques)e(for)h(electronic)f(com-)362 467 y(merce.)30 b(BusinessBots,)20 b(Inc.,)g(a)g(startup)h(compan)o(y)l (,)e(is)h(a)g(mem)o(b)q(er)d(of)j(the)g(Com-)362 528 y(merceNe)o(t)14 b(consortium)h(consisting)h(of)h(o)o(v)o(er)e(500)j (industrial)d(partners.)-38 629 y(2/1997)256 b Fc(Univ)n(ersit)n(y)17 b(of)i(Ronneb)n(y)764 b Fe(Ronneb)o(y)l(,)16 b(Sw)o(eden)362 690 y(In)o(vited)d(consulting)i(on)h(dev)o(eloping)e(a)i(mark)o (et-based)e(m)o(ultiagen)o(t)f(system)h(for)h(electric-)362 750 y(it)o(y)g(distribution)g(for)i(Southern)g(Sw)o(eden)-38 851 y(2/1997)256 b Fc(Lund)18 b(Univ)n(ersit)n(y)989 b Fe(Lund,)17 b(Sw)o(eden)362 912 y(In)o(vited)c(consulting)i(on)h(dev) o(eloping)e(a)i(mark)o(et-based)e(m)o(ultiagen)o(t)f(system)h(for)h (electric-)362 972 y(it)o(y)g(distribution)g(for)i(Southern)g(Sw)o (eden)-38 1073 y(7/1994)256 b Fc(T)-5 b(ec)n(hnical)18 b(Researc)n(h)g(Cen)n(tre)h(of)g(Finland)443 b Fe(Esp)q(o)q(o,)18 b(Finland)362 1134 y Fb(L)n(ab)n(or)n(atory)c(for)j(Information)g(Pr)n (o)n(c)n(essing)362 1194 y Fe(Consulting)f(on)g(further)f(dev)o (elopmen)o(t)e(of)j(the)f(\014elded)g(EPO)h(truc)o(k)f(transp)q (ortation)j(op-)362 1254 y(timization)c(pac)o(k)m(age)-38 1353 y(1993)304 b Fc(T)-5 b(ec)n(hnical)18 b(Researc)n(h)g(Cen)n(tre)h (of)g(Finland)443 b Fe(Esp)q(o)q(o,)18 b(Finland)362 1413 y Fb(L)n(ab)n(or)n(atory)c(for)j(Information)g(Pr)n(o)n(c)n (essing)362 1473 y Fe(Consulting)e(on)h(dev)o(elopmen)o(t)c(of)k(the)e (EPO)i(train)f(transp)q(ortation)i(optimization)d(pac)o(k-)362 1534 y(age)-38 1633 y(1992)304 b Fc(J\177)-28 b(a\177)g(ak\177)g (arinm\177)g(aki)16 b(Co.)893 b Fe(Helsinki,)14 b(Finland)362 1693 y(Executiv)o(e)g(Board)i(Mem)o(b)q(er)-38 1792 y(1989{)280 b Fc(P)n(aro)r(den)n(t)19 b(Co.)997 b Fe(Helsinki,)14 b(Finland)362 1852 y(Executiv)o(e)g(Board)i(Mem)o(b)q(er)-38 1951 y(1989{)280 b Fc(P)n(aro)r(den)n(t)19 b(Co.)997 b Fe(Helsinki,)14 b(Finland)362 2011 y(Buying)h(&)h(con\014guring)h (PCs,)g(p)q(eripherals)f(&)g(soft)o(w)o(are)p eop %%Page: 12 12 12 11 bop -38 150 a Fc(9.)25 b(Publications)p -38 157 371 2 v -38 249 a Fe(Journals)18 b(and)e(highly)g(refereed)f (conferences)22 352 y(1.)24 b(Sandholm,)f(T.)f(1997.)40 b(Unenforced)22 b(Ecommerc)o(e)d(T)l(ransactions.)40 b Fb(IEEE)23 b(Internet)h(Computing)p Fe(,)84 413 y(No)o(v-Dec,)15 b(Sp)q(ecial)h(issue)g(on)h(In)o(ternet)e(Economics.)22 510 y(2.)24 b(Sandholm,)17 b(T.)g(and)h(Lesser,)f(V.)g(1997.)26 b(Coalitions)17 b(among)h(Computationally)e(Bounded)h(Agen)o(ts.)84 570 y Fb(A)o(rti\014cial)j(Intel)r(ligenc)o(e)g Fe(94\(1\),)f(pp.)25 b(99{137,)20 b(Sp)q(ecial)d(issue)h(on)g(Economic)e(Principles)g(of)i (Multia-)84 630 y(gen)o(t)e(Systems.)22 728 y(3.)24 b(Sandholm,)d(T.)g (and)g(Ygge,)h(F.)e(1997.)36 b(On)21 b(the)g(Gains)g(and)g(Losses)h(of) f(Sp)q(eculation)g(in)f(Equilib-)84 788 y(rium)15 b(Mark)o(ets.)22 b Fb(Pr)n(o)n(c)n(e)n(e)n(dings)16 b(of)i(the)g(F)l(ourte)n(enth)g (International)i(Joint)d(Confer)n(enc)n(e)i(on)f(A)o(rti\014cial)84 848 y(Intel)r(ligen)q(c)n(e)j(\(IJCAI-97\))p Fe(,)15 b(pp.)21 b(632{638,)e(Nago)o(y)o(a,)d(Japan.)22 b(\(Acceptance)15 b(rate)h(24\045\))22 945 y(4.)24 b(Sandholm,)c(T.)f(1997.)34 b(Necessary)18 b(and)j(Su\016cien)o(t)d(Con)o(tract)i(T)o(yp)q(es)g (for)g(Optimal)e(T)l(ask)i(Allo)q(ca-)84 1006 y(tion.)34 b Fb(Pr)n(o)n(c)n(e)n(e)n(dings)20 b(of)h(the)h(F)l(ourte)n(enth)g (International)g(Joint)g(Confer)n(enc)n(e)f(on)h(A)o(rti\014cial)g (Intel)r(li-)84 1066 y(genc)n(e)d(\(IJCAI-97\))p Fe(,)d(Nago)o(y)o(a,)g (Japan.)22 b(P)o(oster)16 b(presen)o(tation.)21 b(\(Acceptance)15 b(rate)h(47\045\))22 1163 y(5.)24 b(Sandholm,)13 b(T.)g(1996.)22 b(Limitations)12 b(of)h(the)g(Vic)o(krey)e(Auction)h(in)h (Computational)g(Multiagen)o(t)f(Sys-)84 1223 y(tems.)19 b Fb(Pr)n(o)n(c)n(e)n(e)n(dings)12 b(of)h(the)g(Se)n(c)n(ond)h (International)g(Confer)n(enc)n(e)g(on)g(Multiagent)h(Systems)e (\(ICMAS-)84 1283 y(96\))p Fe(,)j(pp.)21 b(299{306,)e(Keihanna)d (Plaza,)g(Ky)o(oto,)g(Japan,)h(Decem)o(b)q(er.)i(\(Acceptance)c(rate)h (28\045\))22 1381 y(6.)24 b(Sandholm,)d(T.)f(and)h(Lesser,)h(V.)d (1996.)36 b(Adv)m(an)o(tages)21 b(of)f(a)h(Lev)o(eled)e(Commitm)o(en)o (t)e(Con)o(tracting)84 1441 y(Proto)q(col.)46 b Fb(Pr)n(o)n(c)n(e)n(e)n (dings)23 b(of)h(the)h(Thirte)n(enth)g(National)h(Confer)n(enc)n(e)f (on)g(A)o(rti\014cial)g(Intel)r(ligenc)o(e)84 1501 y(\(AAAI-96\))p Fe(,)17 b(pp.)k(126{133,)e(P)o(ortland,)d(OR,)f(August.)22 b(\(Acceptance)15 b(rate)h(31\045\))22 1598 y(7.)24 b(Sandholm,)18 b(T.)f(1996.)28 b(A)17 b(Second)h(Order)g(P)o(arameter)e(for)i(3SA)l (T.)g Fb(Pr)n(o)n(c)n(e)n(e)n(dings)g(of)h(the)g(Thirte)n(enth)84 1659 y(National)24 b(Confer)n(enc)n(e)e(on)h(A)o(rti\014cial)g(Intel)r (ligenc)o(e)i(\(AAAI-96\))p Fe(,)e(pp.)37 b(259{265,)25 b(P)o(ortland,)d(OR,)84 1719 y(August.)g(\(Acceptance)15 b(rate)h(31\045\))22 1816 y(8.)24 b(Sandholm,)d(T.)g(and)g(Crites,)g (R.)f(1995.)36 b(Multiagen)o(t)20 b(Reinforcemen)n(t)e(Learning)j(in)g (the)f(Iterated)84 1876 y(Prisoner's)13 b(Dilemm)o(a.)k Fb(BioSystems)c Fe(37,)g(pp.)20 b(147-166,)c(Sp)q(ecial)11 b(issue)h(on)h(the)f(Prisoner's)g(Dilemma.)22 1974 y(9.)24 b(Sandholm,)13 b(T.)h(and)g(Lesser,)g(V.)e(1995.)22 b(Coalition)14 b(F)l(ormation)f(among)g(Bounded)h(Rational)f(Agen)o(ts.)84 2034 y Fb(Pr)n(o)n(c)n(e)n(e)n(dings)24 b(of)h(the)h(F)l(ourte)n(enth)f (International)i(Joint)e(Confer)n(enc)n(e)h(on)f(A)o(rti\014cial)h (Intel)r(ligenc)o(e)84 2094 y(\(IJCAI-95\))p Fe(,)16 b(pp.)21 b(662{669,)d(Mon)o(treal,)e(Canada.)22 b(\(Acceptance)15 b(rate)h(22\045\))-2 2191 y(10.)24 b(Sandholm,)f(T.)f(and)h(Lesser,)g (V.)f(1995.)40 b(Equilibrium)20 b(Analysis)h(of)i(the)f(P)o (ossibilities)e(of)j(Unen-)84 2252 y(forced)14 b(Exc)o(hange)f(in)h (Multiagen)o(t)e(Systems.)19 b Fb(Pr)n(o)n(c)n(e)n(e)n(dings)14 b(of)h(the)g(F)l(ourte)n(enth)h(International)g(Joint)84 2312 y(Confer)n(enc)n(e)21 b(on)g(A)o(rti\014cial)g(Intel)r(ligenc)o(e) i(\(IJCAI-95\))p Fe(,)c(pp.)31 b(694{701,)22 b(Mon)o(treal,)d(Canada.) 32 b(\(Ac-)84 2372 y(ceptance)16 b(rate)g(22\045\))-2 2469 y(11.)24 b(Sandholm,)14 b(T.)h(and)g(Lesser,)g(V.)e(1995.)23 b(Issues)14 b(in)h(Automated)e(Negotiation)h(and)i(Electronic)d(Com-)84 2530 y(merce:)28 b(Extending)21 b(the)f(Con)o(tract)h(Net)f(F)l(ramew)o (ork.)33 b Fb(Pr)n(o)n(c)n(e)n(e)n(dings)21 b(of)g(the)h(First)f (International)84 2590 y(Confer)n(enc)n(e)f(on)f(Multiagent)h(Systems)f (\(ICMAS-95\))p Fe(,)f(pp.)25 b(328{335,)c(San)d(F)l(rancisco,)f (California.)84 2650 y(\(Acceptance)e(rate)h(33\045\))p eop %%Page: 13 13 13 12 bop -2 150 a Fe(12.)24 b(Neiman,)13 b(D.,)i(Hildum,)d(D.,)j (Lesser,)g(V.)f(and)i(Sandholm,)e(T.)h(1994.)22 b(Exploiting)15 b(Meta-Lev)o(el)e(Infor-)84 210 y(mation)k(in)f(a)i(Distributed)e(Sc)o (heduling)g(System.)23 b Fb(Pr)n(o)n(c)n(e)n(e)n(dings)17 b(of)h(the)g(Twelfth)i(National)f(Confer-)84 270 y(enc)n(e)d(on)f(A)o (rti\014cial)g(Intel)r(ligen)q(c)n(e)j(\(AAAI-94\))p Fe(,)c(pp.)20 b(394{400,)c(Seattle,)d(W)l(ashington.)21 b(\(Acceptance)84 330 y(rate)c(31\045\))-2 432 y(13.)24 b(Sandholm,)15 b(T.)h(1993.)22 b(An)16 b(Implem)o(en)o(tati)o(on)e(of)i (the)g(Con)o(tract)g(Net)f(Proto)q(col)h(Based)g(on)g(Marginal)84 492 y(Cost)j(Calculations.)27 b Fb(Pr)n(o)n(c)n(e)n(e)n(dings)18 b(of)h(the)h(Eleventh)i(National)e(Confer)n(enc)n(e)g(on)f(A)o (rti\014cial)h(Intel)r(li-)84 552 y(genc)n(e)f(\(AAAI-93\))p Fe(,)e(pp.)k(256{262,)e(W)l(ashington)e(DC.)f(\(Acceptance)f(rate)h (24\045\))-38 690 y(Journal)h(articles)f(to)g(b)q(e)h(submitted)35 806 y Fa(\017)24 b Fe(Sandholm,)14 b(T.)h(and)g(Lesser,)g(V.)e(1997.)23 b(Issues)14 b(in)h(Automated)e(Negotiation)h(and)i(Electronic)d(Com-)84 867 y(merce:)26 b(Extending)19 b(the)g(Con)o(tract)h(Net)e(F)l(ramew)o (ork.)29 b(Extended)19 b(v)o(ersion.)30 b Fb(IEEE)20 b(T)l(r)n(ansactions)84 927 y(on)e(Par)n(al)r(lel)h(and)e(Distribute)n (d)h(Systems.)k Fe(In)o(vited)14 b(pap)q(er.)35 1028 y Fa(\017)24 b Fe(Sandholm,)d(T.)f(and)h(Lesser,)h(V.)d(1997.)36 b(Adv)m(an)o(tages)21 b(of)f(a)h(Lev)o(eled)e(Commitm)o(en)o(t)e(Con)o (tracting)84 1089 y(Proto)q(col.)23 b Fb(A)o(rti\014cial)18 b(Intel)r(ligenc)n(e)p Fe(.)35 1190 y Fa(\017)24 b Fe(Sandholm,)14 b(T.)f(and)i(Lesser,)f(V.)f(1997.)22 b(Equilibrium)11 b(Analysis)i(of)h(the)g(P)o(ossibilities)e(of)i(Unenforced)84 1251 y(Exc)o(hange)j(in)e(Multiagen)o(t)h(Systems.)k Fb(A)o(rti\014cial)e(Intel)r(ligenc)o(e)p Fe(.)35 1352 y Fa(\017)24 b Fe(Sandholm,)g(T.)e(and)h(Lesser,)h(V.)e(1997.)42 b(Optimal)20 b(Information)i(V)l(alue)g(Based)g(T)l(ermination)f(of)84 1412 y(An)o(ytime)14 b(Algorithms)g(with)i(Conditional)h(P)o (erformance)d(Pro\014les.)22 b Fb(R)n(e)n(al-Time)17 b(Systems)p Fe(.)35 1514 y Fa(\017)24 b Fe(Sandholm,)15 b(T.,)g(Bro)q(dley)l(,)g(C.,)g(Vido)o(vik,)f(A.,)g(Rita,)h(H.)g(and)h (Sandholm,)f(M.)g(1997.)23 b(Comparison)15 b(of)84 1574 y(Neural)f(Net)o(w)o(orks,)g(Regression)g(Metho)q(ds)h(and)g(Sym)o(b)q (olic)e(Induction)h(Metho)q(ds)h(in)f(Morbidit)o(y)f(and)84 1634 y(Mortalit)o(y)i(Prediction)g(in)h(Equine)f(Gastroin)o(testinal)h (Colic.)k Fb(A)o(rti\014cial)e(Intel)r(ligenc)o(e)i(in)d(Me)n(dicine)p Fe(.)35 1736 y Fa(\017)24 b Fe(Sandholm,)12 b(T.)g(1997.)21 b(TRA)o(CONET:)11 b(An)h(Implem)o(en)n(tation)e(of)i(the)f(Con)o(tract) i(Net)e(Proto)q(col)i(Based)84 1796 y(on)k(Marginal)f(Cost)i (Calculations.)j Fb(Gr)n(oup)16 b(De)n(cision)i(and)f(Ne)n(gotiation)p Fe(.)35 1898 y Fa(\017)24 b Fe(Sandholm,)17 b(T.)g(and)h(Lesser,)f(V.)g (1997.)26 b(Issues)17 b(in)g(Extending)g(the)g(Con)o(tract)h(Net)f(F)l (ramew)o(ork)e(for)84 1958 y(Self-in)o(terested)g(Resource-b)q(ounded)i (Reasoning)g(Agen)o(ts.)k Fb(Gr)n(oup)16 b(De)n(cision)h(and)h(Ne)n (gotiation)p Fe(.)35 2060 y Fa(\017)24 b Fe(Sandholm,)f(T.)f(1997.)41 b(Limitations)21 b(of)h(the)g(Vic)o(krey)e(Auction)i(in)g (Computational)f(Multiagen)o(t)84 2120 y(Systems.)f Fb(IEEE)e(T)l(r)n (ansactions)f(on)h(Systems,)g(Man,)f(and)h(Cyb)n(ernetics)p Fe(.)35 2222 y Fa(\017)24 b Fe(Sandholm,)15 b(T.)h(1997.)22 b(A)16 b(Second)f(Order)h(P)o(arameter)e(for)i(3SA)l(T.)f Fb(Journal)i(of)g(A)o(utomate)n(d)g(R)n(e)n(ason-)84 2282 y(ing)p Fe(.)35 2384 y Fa(\017)24 b Fe(Sandholm,)18 b(T.)f(1997.)28 b(A)18 b(Review)f(of)h(Micro)q(economic)e(F)l (oundations)j(in)e(Computational)h(Multia-)84 2444 y(gen)o(t)e (Systems.)k Fb(A)o(rti\014cial)f(Intel)r(ligenc)n(e)p Fe(.)35 2546 y Fa(\017)24 b Fe(Sandholm,)19 b(T.)f(1997.)30 b(Necessary)18 b(and)h(Su\016cien)o(t)e(Con)o(tract)i(T)o(yp)q(es)g (for)g(Reac)o(hing)f(the)g(Globally)84 2606 y(Optimal)d(T)l(ask)i(Allo) q(cation.)j Fb(A)o(rti\014cial)f(Intel)r(ligenc)n(e)p Fe(.)p eop %%Page: 14 14 14 13 bop -38 150 a Fe(Bo)q(oks)17 b(and)g(b)q(o)q(ok)h(c)o(hapters)-2 259 y(14.)24 b(Sandholm,)14 b(T.)h(and)g(Lesser,)g(V.)e(1997.)23 b(Issues)14 b(in)h(Automated)e(Negotiation)h(and)i(Electronic)d(Com-)84 319 y(merce:)24 b(Extending)19 b(the)f(Con)o(tract)h(Net)f(F)l(ramew)o (ork.)27 b(In)18 b Fb(R)n(e)n(adings)h(in)h(A)n(gents)p Fe(,)g(Huhns,)f(M.)f(and)84 379 y(Singh,)e(M.)g(\(eds.\),)f(Morgan)i (Kaufmann)f(Publishers,)g(Octob)q(er.)-2 478 y(15.)24 b(Sandholm,)19 b(T.)f(1997.)31 b Fb(Ne)n(gotiation)21 b(among)f(Self-Inter)n(este)n(d)i(Computational)r(ly)f(Limite)n(d)e(A)n (gents)p Fe(.)84 539 y(Springer-V)l(erlag.)i(In)o(vited)15 b(b)q(o)q(ok,)i(in)f(preparation.)-2 638 y(16.)24 b(Sandholm,)e(T.)f (1997.)38 b(Distributed)20 b(Rational)i(Decision)f(Making.)36 b(In)21 b(the)g(textb)q(o)q(ok)h Fb(Intr)n(o)n(duc-)84 698 y(tion)f(to)g(Distribute)n(d)f(A)o(rti\014cial)i(Intel)r(ligenc)n (e)p Fe(,)h(W)l(ei\031,)d(G.,)g(ed.,)f(MIT)g(Press.)31 b(In)o(vited)18 b(c)o(hapter,)i(in)84 758 y(preparation.)-2 857 y(17.)k(Sandholm,)g(T.)f(and)h(Crites,)g(R.)e(1996.)44 b(On)23 b(Multiagen)o(t)f(Q-Learning)i(in)f(a)g(Semi-comp)q(etiti)o(v)o (e)84 917 y(Domain.)43 b(In)23 b Fb(A)n(daptation)h(and)g(L)n(e)n (arning)g(in)g(Multi-A)n(gent)i(Systems)p Fe(,)g(W)l(ei\031,)e(G.)f (and)i(Sen,)f(S.,)84 978 y(eds.)k(Lecture)18 b(Notes)g(in)g (Arti\014cial)e(In)o(telligence)g(1042)k(of)e(Lecture)g(Notes)g(in)g (Computer)f(Science,)84 1038 y(Springer-V)l(erlag,)f(pp.)21 b(191{205.)-2 1137 y(18.)j(Honk)o(ela,)18 b(T.)f(and)i(Sandholm,)f(T.)f (1993.)29 b(Mac)o(hine)17 b(Learning.)27 b(In)18 b(the)g Fb(Finnish)h(Encyclop)n(e)n(dia)h(of)84 1197 y(A)o(rti\014cial)f(Intel) r(ligenc)n(e)p Fe(,)g(Eero)e(Hyv\177)-24 b(onen,)15 b(ed.,)g(pp.)21 b(244{255.)-2 1296 y(19.)j(Sandholm,)c(T.)g(1991.)33 b(Solving)20 b(the)f(1-Dimensional)g(F)l(ractal)g(In)o(v)o(erse)f (Problem)h(with)g(a)i(Genetic)84 1356 y(Algorithm.)26 b(In)18 b Fb(Genetic)i(A)o(lgorithms)p Fe(,)f(Jarmo)f(T.)g(Alander,)f (ed.,)h(pp.)27 b(126{132,)21 b(Publications)d(of)84 1417 y(the)e(Helsinki)f(Univ)o(ersit)o(y)e(of)k(T)l(ec)o(hnology)l(.)-2 1516 y(20.)24 b(Aho,)19 b(J.,)f(Arpp)q(e,)g(A.,)f(Henrikson,)h(M.,)f (Hin)o(tsa,)h(J.,)g(Hypp\177)-24 b(onen,)18 b(A.,)f(Jouppi,)i(J)f(.,)g (Linnanen,)h(L.,)84 1576 y(Luh)o(tala,)f(M.,)f(L\177)-24 b(a\177)g(ar\177)g(a,)19 b(M.,)d(Matero,)h(S.,)g(P)o(ank)m(ak)o(oski,)g (J.,)g(Pulli,)f(V.,)h(Rahik)m(ainen,)g(S.,)g(Sandholm,)84 1636 y(T.,)f(eds.,)f(1991.)23 b Fb(The)18 b(Innovative)h(Enterprise)p Fe(.)j(Pro)q(dek)o(o)16 b(Publications,)g(Esp)q(o)q(o,)i(Finland.)-38 1801 y(Other)e(refereed)f(conferences)-2 1910 y(21.)24 b(Sandholm,)c(T.)g(1998.)35 b(Coalition)20 b(F)l(ormation)f(under)h (Costly)h(Computation.)32 b Fb(Institute)22 b(for)f(Op-)84 1970 y(er)n(ations)h(R)n(ese)n(ar)n(ch)f(and)h(the)h(Management)h (Scienc)n(es)g(\(INF)o(ORMS\))d(International)j(c)n(onfer)n(enc)n(e,)84 2030 y(Col)r(le)n(ge)d(on)f(Gr)n(oup)d(De)n(cision)i(and)h(Ne)n (gotiation,)g(Game)f(The)n(ory)e(and)j(Applic)n(ations)f(tr)n(ack)p Fe(,)f(Mon-)84 2091 y(treal,)e(Canada.)23 b(In)o(vited)14 b(abstract.)-2 2190 y(22.)24 b(Sandholm,)e(M.,)g(Sandholm.)37 b(T.,)23 b(Bro)q(dley)l(,)f(C.,)g(and)g(Vido)o(vic,)f(A.)g(1996.)39 b(Linear)22 b(and)g(logistic)84 2250 y(regression,)c(sym)o(b)q(olic)e (induction)h(metho)q(ds,)h(and)g(neural)g(net)o(w)o(orks)f(in)h (morbidit)o(y)d(diagnosis)k(and)84 2310 y(mortalit)o(y)e(prediction)h (in)h(equine)f(gastroin)o(testinal)h(colic:)25 b(An)19 b(extended)f(abstract.)30 b Fb(The)20 b(Se)n(c)n(ond)84 2370 y(A)o(nnual)e(SEPSIS/SIRS)f(Confer)n(enc)n(e:)23 b(R)n(e)n(ducing)17 b(Mortality)e(to)i(Patients)g(&)g(Suppliers)p Fe(,)e(W)l(ashing-)84 2430 y(ton,)i(D.C.)f(P)o(oster)g(presen)o (tation.)-2 2530 y(23.)24 b(Sandholm,)16 b(T.)h(and)g(Lesser,)g(V.)f (1995.)24 b(On)17 b(Automated)f(Con)o(tracting)h(in)g(Multi-en)o (terprise)d(Man-)84 2590 y(ufacturing.)34 b Fb(Pr)n(o)n(c)n(e)n(e)n (dings)20 b(of)h(the)h(\\Impr)n(oving)f(Manufacturing)h(Performanc)n(e) f(in)g(a)h(Distribute)n(d)84 2650 y(Enterprise:)h(A)n(dvanc)n(e)n(d)18 b(Systems)f(and)h(T)l(o)n(ols")f(Confer)n(enc)n(e)p Fe(,)g(pp.)k (33{42,)d(Edin)o(burgh,)e(Scotland.)p eop %%Page: 15 15 15 14 bop -2 150 a Fe(24.)24 b(Sandholm,)h(T.)f(1992.)46 b(Automatic)22 b(Co)q(op)q(eration)k(of)e(Area-Distributed)f(Dispatc)o (h)h(Cen)o(ters)g(in)84 210 y(V)l(ehicle)d(Routing.)40 b Fb(Pr)n(o)n(c)n(e)n(e)n(dings)22 b(of)h(the)g(International)i(Confer) n(enc)n(e)f(on)f(A)o(rti\014cial)h(Intel)r(ligenc)o(e)84 270 y(Applic)n(ations)18 b(in)g(T)l(r)n(ansp)n(ortation)f(Engine)n (ering)p Fe(,)g(pp.)22 b(449{467,)c(San)f(Buena)o(v)o(en)o(tura,)d (California.)-2 372 y(25.)24 b(Sandholm,)12 b(T.)g(1992.)21 b(Automatic)11 b(Co)q(op)q(eration)j(of)e(F)l(actorially)f(Distributed) g(Dispatc)o(h)h(Cen)o(ters)g(in)84 432 y(V)l(ehicle)g(Routing.)21 b Fb(A)o(bstr)n(act)15 b(Col)r(le)n(ction)i(of)d(the)i(Joint)f (International)h(Confer)n(enc)n(e)g(on)g(Op)n(er)n(ational)84 492 y(R)n(ese)n(ar)n(ch)g(/)i(Management)h(Scienc)n(e)g(\(EUR)o(O)e(/)h (TIMS)f(-92\))p Fe(,)f(Helsinki,)e(Finland.)-2 594 y(26.)24 b(Sandholm,)g(T.)e(1992.)43 b(A)22 b(Bargaining)h(Net)o(w)o(ork)f(for)h (In)o(telligen)o(t)d(Agen)o(ts.)40 b Fb(Pr)n(o)n(c)n(e)n(e)n(dings)23 b(of)g(the)84 654 y(Finnish)18 b(A)o(rti\014cial)h(Intel)r(ligenc)n(e)i (Confer)n(enc)n(e)d(\(ST)l(eP-92\),)h(New)g(Dir)n(e)n(ctions)e(in)h(A)o (rti\014cial)g(Intel)r(li-)84 714 y(genc)n(e)p Fe(,)g(V)l(ol.)i(3,)d (pp.)k(173{181,)e(Esp)q(o)q(o,)e(Finland.)-2 816 y(27.)24 b(Linnainmaa,)h(S.,)g(Jokinen,)g(O.,)g(Sandholm,)f(T.)g(and)g(V)l (epsalainen,)h(A.)e(M.)g(1992.)45 b(Adv)m(anced)84 876 y(Computer)18 b(Supp)q(orted)i(V)l(ehicle)d(Routing)i(for)g(Hea)o(vy)f (T)l(ransp)q(orts.)30 b Fb(Pr)n(o)n(c)n(e)n(e)n(dings)19 b(of)h(the)g(Finnish)84 936 y(A)o(rti\014cial)15 b(Intel)r(ligenc)o(e)i (Confer)n(enc)n(e)d(\(ST)l(eP-92\),)i(New)f(Dir)n(e)n(ctions)e(in)h(A)o (rti\014cial)h(Intel)r(ligenc)o(e)p Fe(,)h(V)l(ol.)84 997 y(3,)h(pp.)k(163{172,)e(Esp)q(o)q(o,)e(Finland.)-2 1098 y(28.)24 b(Sandholm,)d(T.)g(1991.)36 b(A)20 b(Strategy)h(for)g (Decreasing)f(the)h(T)l(otal)g(T)l(ransp)q(ortation)i(Costs)e(Among)84 1158 y(Area-Distributed)14 b(T)l(ransp)q(ortation)i(Cen)o(ters.)k Fb(Pr)n(o)n(c)n(e)n(e)n(dings)15 b(of)g(the)h(\\Nor)n(dic)f(Op)n(er)n (ations)g(A)o(nalysis)84 1219 y(in)j(Co)n(op)n(er)n(ation)e(\(NO)o (AS-91\):)23 b(OR)18 b(in)g(Business")g(Confer)n(enc)n(e)p Fe(,)e(T)l(urku,)g(Finland.)-38 1391 y(Refereed)f(w)o(orkshops)-2 1508 y(29.)24 b(Sandholm,)16 b(T.)g(1998.)24 b(Negotiation)17 b(among)f(Computationally)g(Limited)f(Self-In)o(terested)f(Agen)o(ts.) 84 1568 y Fb(Se)n(c)n(ond)j(International)g(Workshop)e(on)h(Co)n(op)n (er)n(ative)f(Information)g(A)n(gents)i(\(CIA\):)f(L)n(e)n(arning,)f (Mo-)84 1628 y(bility)j(and)f(Ele)n(ctr)n(onic)h(Commer)n(c)n(e)f(for)f (Information)h(Disc)n(overy)g(in)g(the)h(Internet)p Fe(.)k(In)o(vited) 14 b(pap)q(er.)84 1689 y(Cite)i(de)g(Sciences)f(-)i(La)g(Vilette,)c(P)o (aris,)j(F)l(rance,)g(July)f(3-8.)-2 1790 y(30.)24 b(T)l(ohme,)19 b(F.)f(and)i(Sandholm,)f(T.)g(1997.)32 b(Coalition)19 b(F)l(ormation)f(Pro)q(cesses)i(with)f(Belief)e(Revision)84 1850 y(among)d(Bounded)f(Rational)g(Self-In)o(terested)e(Agen)o(ts.)20 b(Fifteen)o(th)12 b(In)o(ternational)g(Join)o(t)h(Conference)84 1911 y(on)g(Arti\014cial)d(In)o(telligence)f(\(IJCAI-97\),)k(W)l (orkshop)g(on)f(So)q(cial)g(In)o(teraction)f(and)h(Comm)o(unit)o(yw)o (are,)84 1971 y(Nago)o(y)o(a,)k(Japan,)h(August)g(25.)-2 2073 y(31.)24 b(Sandholm.)19 b(T.,)12 b(Bro)q(dley)l(,)f(C.,)h(Vido)o (vic,)e(A.)g(and)i(Sandholm,)g(M.)e(1996.)21 b(Comparison)12 b(of)f(Regression)84 2133 y(Metho)q(ds,)17 b(Sym)o(b)q(olic)e (Induction)h(Metho)q(ds)i(and)f(Neural)f(Net)o(w)o(orks)g(in)h (Morbidit)o(y)e(Diagnosis)j(and)84 2193 y(Mortalit)o(y)e(Prediction)g (in)h(Equine)f(Gastroin)o(testinal)h(Colic.)22 b Fb(Working)c(Notes)h (of)f(the)g(AAAI)h(1996)84 2253 y(Spring)26 b(Symp)n(osium)e(Series,)j (A)o(rti\014cial)f(Intel)r(ligenc)o(e)i(in)d(Me)n(dicine:)38 b(Applic)n(ations)25 b(of)g(Curr)n(ent)84 2313 y(T)l(e)n(chnolo)n(gies) p Fe(,)18 b(pp.)j(154{159,)d(Stanford)g(Univ)o(ersit)o(y)l(,)13 b(California.)-2 2415 y(32.)24 b(Sandholm,)g(T.)f(and)h(Crites,)g(R.)e (1995.)44 b(On)23 b(Multiagen)o(t)f(Q-Learning)i(in)f(a)g(Semi-comp)q (etiti)o(v)o(e)84 2475 y(Domain.)34 b Fb(Working)22 b(Notes)g(of)f(the) h(F)l(ourte)n(enth)g(International)h(Joint)f(Confer)n(enc)n(e)g(on)g(A) o(rti\014cial)84 2535 y(Intel)r(ligen)q(c)n(e)k(\(IJCAI-95\))d (Workshop)f(on)i(A)n(daptation)f(and)g(L)n(e)n(arning)g(in)h (Multiagent)h(Systems)p Fe(,)84 2596 y(pp.)d(71{77,)c(Mon)o(treal,)d (Canada.)p eop %%Page: 16 16 16 15 bop -2 150 a Fe(33.)24 b(Sandholm,)d(T.)f(and)h(Lesser,)h(V.)d (1994.)36 b(An)20 b(Exc)o(hange)g(Proto)q(col)h(without)g(Enforcemen)o (t.)32 b Fb(Pr)n(o-)84 210 y(c)n(e)n(e)n(dings)23 b(of)g(the)g(Thirte)n (enth)g(International)h(Workshop)d(on)i(Distribute)n(d)g(A)o (rti\014cial)g(Intel)r(ligenc)o(e)84 270 y(\(D)o(AI-94\))p Fe(,)16 b(pp.)21 b(305{319,)e(Seattle,)c(W)l(ashington.)-2 372 y(34.)24 b(Sandholm,)12 b(T.)f(1994.)21 b(A)11 b(New)g(Order)g(P)o (arameter)f(for)i(3SA)l(T.)f Fb(Pr)n(o)n(c)n(e)n(e)n(dings)h(of)h(the)g (Twelfth)i(National)84 432 y(Confer)n(enc)n(e)i(on)f(A)o(rti\014cial)g (Intel)r(ligen)q(c)n(e)j(\(AAAI-94\))d(Workshop)f(on)h(Exp)n(erimental) h(Evaluation)g(of)84 492 y(R)n(e)n(asoning)h(and)f(Se)n(ar)n(ch)g (Metho)n(ds)p Fe(,)f(pp.)21 b(57{63,)d(Seattle,)d(W)l(ashington.)-2 594 y(35.)24 b(Sandholm,)e(T.)e(and)i(Lesser,)g(V.)e(1994.)37 b(Utilit)o(y-Based)19 b(T)l(ermination)g(of)j(An)o(ytime)c(Algorithms.) 84 654 y Fb(Pr)n(o)n(c)n(e)n(e)n(dings)k(of)g(the)h(Eur)n(op)n(e)n(an)e (Confer)n(enc)n(e)i(on)g(A)o(rti\014cial)g(Intel)r(ligenc)o(e)i (\(ECAI-94\))d(Workshop)84 714 y(on)c(De)n(cision)g(The)n(ory)e(for)h (D)o(AI)g(Applic)n(ations)p Fe(,)f(pp.)22 b(88{99,)17 b(Amsterdam,)c(The)k(Netherlands.)-2 816 y(36.)24 b(Sandholm,)12 b(T.)g(1994.)21 b(Automatic)10 b(Co)q(op)q(eration)k(of)e (Area-Distributed)f(Dispatc)o(h)h(Cen)o(ters)f(in)g(V)l(ehi-)84 876 y(cle)i(Routing.)21 b Fb(Or)n(ganization)15 b(for)g(Ec)n(onomic)g (Co)n(op)n(er)n(ation)f(and)h(Development)i(\(OECD\))e(Scienti\014c)84 936 y(Exp)n(ert)j(Gr)n(oup)e(TT6)h(me)n(eting)i(on)f(A)n(dvanc)n(e)n(d) f(L)n(o)n(gistics)g(and)g(Information)h(T)l(e)n(chnolo)n(gy)g(in)g(F)l (r)n(eight)84 997 y(T)l(r)n(ansp)n(ort)p Fe(,)d(W)l(ashington,)i(D.C.) -2 1098 y(37.)24 b(Sandholm,)15 b(T.)h(1993.)22 b(An)16 b(Implem)o(en)o(tati)o(on)e(of)i(the)g(Con)o(tract)g(Net)f(Proto)q(col) h(Based)g(on)g(Marginal)84 1158 y(Cost)h(Calculations.)22 b Fb(Pr)n(o)n(c)n(e)n(e)n(dings)16 b(of)h(the)h(Twelfth)g (International)h(Workshop)e(on)g(Distribute)n(d)h(A)o(r-)84 1219 y(ti\014cial)h(Intel)r(ligenc)o(e)h(\(D)o(AI-93\))p Fe(,)c(pp.)21 b(295{308,)e(Hidden)c(V)l(alley)l(,)f(P)o(ennsylv)m (ania.)-38 1391 y(T)l(ec)o(hnical)h(rep)q(orts)-2 1508 y(38.)24 b(Sandholm.)19 b(T.,)12 b(Bro)q(dley)l(,)f(C.,)h(Vido)o(vic,)e (A.)g(and)i(Sandholm,)g(M.)e(1997.)21 b(Comparison)12 b(of)f(Regression)84 1568 y(Metho)q(ds,)17 b(Sym)o(b)q(olic)e (Induction)h(Metho)q(ds)i(and)f(Neural)f(Net)o(w)o(orks)g(in)h (Morbidit)o(y)e(Diagnosis)j(and)84 1628 y(Mortalit)o(y)13 b(Prediction)f(in)h(Equine)f(Gastroin)o(testinal)i(Colic.)19 b Fb(A)o(meric)n(an)c(Asso)n(ciation)g(for)f(A)o(rti\014cial)84 1689 y(Intel)r(ligen)q(c)n(e)21 b(T)l(e)n(chnic)n(al)d(R)n(ep)n(ort)p Fe(,)c(accepted)i(for)g(publication)g(\(refereed\),)e(to)j(app)q(ear.) -2 1790 y(39.)24 b(Sandholm,)d(T.)f(and)h(Lesser,)h(V.)d(1995.)36 b(Adv)m(an)o(tages)21 b(of)f(a)h(Lev)o(eled)e(Commitm)o(en)o(t)e(Con)o (tracting)84 1850 y(Proto)q(col.)27 b(Extended)17 b(v)o(ersion.)24 b Fb(University)c(of)e(Massachusetts)i(at)e(A)o(mherst,)h(Computer)g (Scienc)n(e)84 1911 y(T)l(e)n(chnic)n(al)g(R)n(ep)n(ort)d(TR)h(95-72)p Fe(.)-2 2012 y(40.)24 b(Sandholm,)13 b(T.)h(and)g(Lesser,)g(V.)e(1995.) 22 b(Coalition)14 b(F)l(ormation)f(among)g(Bounded)h(Rational)f(Agen)o (ts.)84 2073 y(Extended)18 b(v)o(ersion.)26 b Fb(University)19 b(of)g(Massachusetts)h(at)f(A)o(mherst,)g(Computer)g(Scienc)n(e)h(T)l (e)n(chnic)n(al)84 2133 y(R)n(ep)n(ort)c(TR)h(95-71)p Fe(.)-2 2234 y(41.)24 b(Berkman,)18 b(N.)g(and)i(Sandholm,)f(T.)g (1995.)31 b(What)20 b(should)g(b)q(e)f(minimi)o(ze)o(d)e(in)i(a)g (decision)g(tree:)26 b(A)84 2295 y(re-examination.)41 b Fb(University)25 b(of)e(Massachusetts)i(at)f(A)o(mherst,)h(Computer)f (Scienc)n(e)h(T)l(e)n(chnic)n(al)84 2355 y(R)n(ep)n(ort)16 b(TR)h(95-20)p Fe(.)-2 2457 y(42.)24 b(Sandholm,)e(T.)e(and)i(Lesser,)g (V.)e(1994.)37 b(Utilit)o(y-Based)19 b(T)l(ermination)g(of)j(An)o (ytime)c(Algorithms.)84 2517 y(Extended)g(v)o(ersion.)26 b Fb(University)19 b(of)g(Massachusetts)h(at)f(A)o(mherst,)g(Computer)g (Scienc)n(e)h(T)l(e)n(chnic)n(al)84 2577 y(R)n(ep)n(ort)c(TR)h(94-54)p Fe(.)p eop %%Page: 17 17 17 16 bop -2 150 a Fe(43.)24 b(Sandholm,)12 b(T.)f(and)h(Lesser,)h(V.)d (1994.)21 b(An)12 b(Exc)o(hange)f(Proto)q(col)i(without)e(Enforcemen)o (t.)18 b(Extended)84 210 y(v)o(ersion.)j Fb(University)c(of)f (Massachusetts)h(at)g(A)o(mherst,)g(Computer)f(Scienc)n(e)i(T)l(e)n (chnic)n(al)g(R)n(ep)n(ort)d(TR)84 270 y(94-44)p Fe(.)-2 372 y(44.)24 b(Sandholm,)16 b(T.)g(1991.)23 b(Automatic)14 b(Co)q(op)q(eration)k(of)f(Dispatc)o(h)f(Cen)o(ters)g(in)g(V)l(ehicle)e (Routing.)22 b Fb(R)n(e-)84 432 y(se)n(ar)n(ch)16 b(R)n(ep)n(ort)g(No.) 22 b(J-9,)16 b(L)n(ab)n(or)n(atory)e(for)i(Information)h(Pr)n(o)n(c)n (essing,)f(T)l(e)n(chnic)n(al)i(R)n(ese)n(ar)n(ch)d(Centr)n(e)84 492 y(of)j(Finland)p Fe(.)-38 665 y(Other)e(publications)-2 782 y(45.)24 b(Sandholm,)16 b(T.)g(1996.)24 b(Negotiation)17 b(among)f(Self-In)o(terested)f(Computationally)h(Limited)e(Agen)o(ts.) 84 842 y Fb(Ph.D.)22 b(Dissertation.)35 b(University)23 b(of)e(Massachusetts)i(at)f(A)o(mherst,)g(Dep)n(artment)g(of)g (Computer)84 902 y(Scienc)n(e)p Fe(.)i(297)17 b(pages.)-2 1004 y(46.)24 b(Sandholm,)18 b(T.)h(1996.)29 b(A)19 b(Ph.D.)f (Dissertation)h(Prop)q(osal)h(for)e(the)h(Study)f(of)h(Negotiation)g (among)84 1064 y(Self-In)o(terested)12 b(Computationally)g(Limited)f (Agen)o(ts.)19 b(Univ)o(ersit)o(y)11 b(of)i(Massac)o(h)o(usetts)g(at)h (Amherst,)84 1124 y(Departmen)o(t)h(of)i(Computer)e(Science.)20 b(119)d(pages.)-2 1226 y(47.)24 b(Sandholm,)d(T.)f(1994.)34 b(A)20 b(New)g(Order)f(P)o(arameter)g(for)i(3SA)l(T)f(/)h(Utilit)o (y-Based)d(T)l(ermination)g(of)84 1286 y(An)o(ytime)f(Algorithms.)29 b(\(2)20 b(separate)f(pap)q(ers\).)32 b Fb(M.S.)20 b(Thesis,)h (University)g(of)g(Massachusetts)f(at)84 1346 y(A)o(mherst,)e(Dep)n (artment)f(of)g(Computer)h(Scienc)n(e)p Fe(.)-2 1448 y(48.)24 b(Sandholm,)18 b(T.)g(1992.)28 b(In)o(tro)q(duction)18 b(to)g(the)g(Logistics)h(W)l(orkshop.)28 b Fb(Pr)n(o)n(c)n(e)n(e)n (dings)18 b(of)h(the)g(Finnish)84 1508 y(A)o(rti\014cial)c(Intel)r (ligenc)o(e)i(Confer)n(enc)n(e)d(\(ST)l(eP-92\),)i(New)f(Dir)n(e)n (ctions)e(in)h(A)o(rti\014cial)h(Intel)r(ligenc)o(e)p Fe(,)h(V)l(ol.)84 1568 y(3,)h(pp.)k(147-148,)d(Esp)q(o)q(o,)g(Finland.) -2 1670 y(49.)24 b(Sandholm,)h(T.)f(1991.)46 b(Automatic)22 b(Co)q(op)q(eration)k(of)f(Dispatc)o(h)f(Cen)o(ters)f(in)h(V)l(ehicle)e (Routing.)84 1730 y Fb(M.S.)d(Thesis,)f(Helsinki)j(University)e(of)f(T) l(e)n(chnolo)n(gy,)h(Industrial)g(Engine)n(ering)h(and)f(Management)84 1790 y(Scienc)n(e)p Fe(,)f(Esp)q(o)q(o,)g(Finland.)p eop %%Page: 18 18 18 17 bop -38 150 a Fc(10.)25 b(Presen)n(tations)p -38 157 429 2 v -38 251 a(a.)g(Conferences)18 b(and)i(w)n(orkshops)f (\(including)f(in)n(vited)f(ones\))22 349 y Fe(1.)24 b(Negotiation)14 b(among)f(Computationally)f(Limited)f(Self-In)o (terested)g(Agen)o(ts.)20 b Fb(Se)n(c)n(ond)15 b(International)84 409 y(Workshop)21 b(on)g(Co)n(op)n(er)n(ative)f(Information)g(A)n (gents)i(\(CIA\):)f(L)n(e)n(arning,)g(Mobility)h(and)f(Ele)n(ctr)n (onic)84 469 y(Commer)n(c)n(e)e(for)h(Information)f(Disc)n(overy)g(in)i (the)f(Internet)p Fe(.)30 b(F)l(unded)19 b(in)o(vited)e(talk.)28 b(Cite)18 b(de)h(Sci-)84 529 y(ences)d(-)h(La)g(Vilette,)c(P)o(aris,)j (F)l(rance,)g(July)f(3-8,)i(1998.)22 624 y(2.)24 b(Negotiation)e(among) f(Computationally)g(Limited)e(Self-In)o(terested)h(Agen)o(ts.)36 b Fb(A)n(gents,)24 b(A)o(life)f(and)84 685 y(Computational)15 b(Ec)n(onomics)f(Workshop)p Fe(.)19 b(F)l(unded)12 b(in)o(vited)e (talk.)20 b(Hewlett)11 b(P)o(ac)o(k)m(ard)h(Labs,)i(Bristol,)84 745 y(UK,)i(June)g(24-25,)i(1998.)22 840 y(3.)24 b(Coalition)13 b(F)l(ormation)f(under)g(Costly)h(Computation.)19 b Fb(Institute)c(for) f(Op)n(er)n(ations)f(R)n(ese)n(ar)n(ch)g(and)h(the)84 900 y(Management)21 b(Scienc)n(es)g(\(INF)o(ORMS\))e(International)i(c) n(onfer)n(enc)n(e,)f(Col)r(le)n(ge)i(on)d(Gr)n(oup)f(De)n(cision)84 960 y(and)i(Ne)n(gotiation,)h(Game)e(The)n(ory)f(and)i(Applic)n(ations) f(tr)n(ack)p Fe(.)28 b(In)o(vited)17 b(talk.)27 b(Mon)o(treal,)17 b(Canada,)84 1021 y(April)e(26-29,)j(1998.)22 1116 y(4.)24 b(Using)14 b(Economics)f(for)h(Resource)f(Allo)q(cation)g(and)i(Con)o (trol.)20 b(Second)14 b(In)o(ternational)f(W)l(orkshop)i(on)84 1176 y(Mobile)f(Agen)o(ts.)21 b(F)l(unded)15 b(in)o(vited)e(talk.)20 b(Dartmouth)15 b(College,)g(Hano)o(v)o(er,)f(NH,)f(Septem)o(b)q(er)g (19-20,)84 1236 y(1997.)22 1331 y(5.)24 b(On)15 b(the)g(Gains)g(and)h (Losses)g(of)f(Sp)q(eculation)g(in)g(Equilibrium)d(Mark)o(ets.)20 b Fb(Pr)n(o)n(c)n(e)n(e)n(dings)15 b(of)h(the)g(F)l(our-)84 1392 y(te)n(enth)21 b(International)f(Joint)f(Confer)n(enc)n(e)h(on)g (A)o(rti\014cial)g(Intel)r(ligenc)o(e)i(\(IJCAI\))p Fe(,)17 b(Nago)o(y)o(a,)h(Japan,)84 1452 y(1997.)23 b(\(Acceptance)15 b(rate)h(24\045\))22 1547 y(6.)24 b(Coalition)11 b(F)l(ormation)f(Pro)q (cesses)i(with)f(Belief)d(Revision)i(among)h(Bounded)g(Rational)g (Self-In)o(terested)84 1607 y(Agen)o(ts.)21 b(Fifteen)o(th)12 b(In)o(ternational)i(Join)o(t)f(Conference)h(on)h(Arti\014cial)d(In)o (telligence)f(\(IJCAI\),)i(W)l(ork-)84 1667 y(shop)18 b(on)e(So)q(cial)h(In)o(teraction)e(and)i(Comm)o(unit)o(y)o(w)o(are,)c (Nago)o(y)o(a,)j(Japan,)h(August)f(25,)h(1997.)22 1762 y(7.)24 b(Limitations)c(of)g(the)g(Vic)o(krey)e(Auction)i(in)g (Computational)g(Multiagen)o(t)f(Systems.)33 b Fb(Pr)n(o)n(c)n(e)n(e)n (dings)84 1823 y(of)22 b(the)g(Se)n(c)n(ond)g(International)h(Confer)n (enc)n(e)f(on)g(Multiagent)h(Systems)f(\(ICMAS\))p Fe(,)e(pp.)34 b(299{306,)84 1883 y(Keihanna)17 b(Plaza,)f(Ky)o(oto,)g(Japan,)h(Decem) o(b)q(er,)c(1996.)23 b(\(Acceptance)15 b(rate)h(28\045\))22 1978 y(8.)24 b(Adv)m(an)o(tages)g(of)e(a)h(Lev)o(eled)f(Commitm)o(e)o (n)o(t)e(Con)o(tracting)j(Proto)q(col.)41 b Fb(Pr)n(o)n(c)n(e)n(e)n (dings)22 b(of)h(the)h(Thir-)84 2038 y(te)n(enth)17 b(National)f (Confer)n(enc)n(e)g(on)g(A)o(rti\014cial)g(Intel)r(ligen)q(c)n(e)i (\(AAAI\))p Fe(,)c(pp.)21 b(126{133,)c(P)o(ortland,)d(OR,)84 2098 y(August,)j(1996.)22 b(\(Acceptance)15 b(rate)h(31\045\))22 2194 y(9.)24 b(A)17 b(Second)h(Order)e(P)o(arameter)g(for)i(3SA)l(T.)f Fb(Pr)n(o)n(c)n(e)n(e)n(dings)g(of)h(the)h(Thirte)n(enth)g(National)g (Confer)n(enc)n(e)84 2254 y(on)d(A)o(rti\014cial)h(Intel)r(ligenc)n(e)i (\(AAAI\))p Fe(,)14 b(pp.)21 b(259{265,)16 b(P)o(ortland,)f(OR,)e (August,)i(1996.)22 b(\(Acceptance)84 2314 y(rate)17 b(31\045\))-2 2409 y(10.)24 b(Comparison)15 b(of)f(Regression)g(Metho)q (ds,)h(Sym)o(b)q(olic)d(Induction)i(Metho)q(ds)g(and)h(Neural)f(Net)o (w)o(orks)f(in)84 2469 y(Morbidit)o(y)j(Diagnosis)h(and)g(Mortalit)o(y) e(Prediction)h(in)g(Equine)g(Gastroin)o(testinal)g(Colic.)21 b Fb(Working)84 2530 y(Notes)g(of)e(the)h(AAAI)h(Spring)f(Symp)n(osium) f(Series,)h(A)o(rti\014cial)h(Intel)r(ligenc)o(e)h(in)e(Me)n(dicine:)28 b(Appli-)84 2590 y(c)n(ations)c(of)g(Curr)n(ent)f(T)l(e)n(chnolo)n (gies)p Fe(,)j(pp.)42 b(154{159,)27 b(Stanford)d(Univ)o(ersit)o(y)l(,)d (California,)j(Marc)o(h,)84 2650 y(1996.)p eop %%Page: 19 19 19 18 bop -2 150 a Fe(11.)24 b(Coalition)h(F)l(ormation)e(among)i (Bounded)f(Rational)h(Agen)o(ts.)45 b Fb(Pr)n(o)n(c)n(e)n(e)n(dings)24 b(of)g(the)i(F)l(ourte)n(enth)84 210 y(International)19 b(Joint)d(Confer)n(enc)n(e)i(on)f(A)o(rti\014cial)g(Intel)r(ligenc)o(e) j(\(IJCAI\))p Fe(,)14 b(pp.)21 b(662{669,)d(Mon)o(treal,)84 270 y(Canada,)g(August,)e(1995.)23 b(\(Acceptance)14 b(rate)j(22\045\))-2 372 y(12.)24 b(Equilibrium)14 b(Analysis)j(of)g (the)f(P)o(ossibilities)g(of)h(Unenforced)f(Exc)o(hange)h(in)f (Multiagen)o(t)g(Systems.)84 432 y Fb(Pr)n(o)n(c)n(e)n(e)n(dings)e(of)g (the)h(F)l(ourte)n(enth)g(International)h(Joint)e(Confer)n(enc)n(e)h (on)g(A)o(rti\014cial)g(Intel)r(ligenc)o(e)i(\(IJ-)84 492 y(CAI\))p Fe(,)f(pp.)21 b(694{701,)e(Mon)o(treal,)c(Canada,)j (August,)e(1995.)22 b(\(Acceptance)15 b(rate)h(22\045\))-2 594 y(13.)24 b(On)c(Multiagen)o(t)e(Q-Learning)j(in)e(a)g(Semi-comp)q (etiti)o(v)o(e)d(Domain.)30 b Fb(Working)20 b(Notes)h(of)g(the)f(F)l (our-)84 654 y(te)n(enth)27 b(International)h(Joint)e(Confer)n(enc)n(e) g(on)h(A)o(rti\014cial)f(Intel)r(ligen)q(c)n(e)j(\(IJCAI\))c(Workshop)g (on)84 714 y(A)n(daptation)20 b(and)g(L)n(e)n(arning)f(in)h(Multiagent) h(Systems)p Fe(,)e(pp.)29 b(71{77,)20 b(Mon)o(treal,)e(Canada,)j (August,)84 774 y(1995.)-2 876 y(14.)j(Issues)17 b(in)e(Automated)g (Negotiation)h(and)h(Electronic)e(Commerce:)j(Extending)e(the)g(Con)o (tract)g(Net)84 936 y(F)l(ramew)o(ork.)43 b Fb(Pr)n(o)n(c)n(e)n(e)n (dings)24 b(of)g(the)i(First)e(International)i(Confer)n(enc)n(e)f(on)g (Multiagent)h(Systems)84 997 y(\(ICMAS\))p Fe(,)16 b(pp.)21 b(328{335,)e(San)e(F)l(rancisco,)e(California,)h(June,)g(1995.)23 b(\(Acceptance)14 b(rate)j(33\045\))-2 1098 y(15.)24 b(On)c(Automated)e(Con)o(tracting)i(in)f(Multi-en)o(terprise)e(Man)o (ufacturing.)30 b Fb(Pr)n(o)n(c)n(e)n(e)n(dings)20 b(of)g(the)h(\\Im-) 84 1158 y(pr)n(oving)f(Manufacturing)h(Performanc)n(e)e(in)h(a)f (Distribute)n(d)h(Enterprise:)27 b(A)n(dvanc)n(e)n(d)19 b(Systems)h(and)84 1219 y(T)l(o)n(ols")e(Confer)n(enc)n(e)p Fe(,)e(pp.)22 b(33{42,)17 b(Edin)o(burgh,)f(Scotland,)g(July)l(,)g (1995.)-2 1320 y(16.)24 b(An)17 b(Exc)o(hange)f(Proto)q(col)i(without)e (Enforcemen)o(t.)21 b Fb(Pr)n(o)n(c)n(e)n(e)n(dings)16 b(of)i(the)g(Thirte)n(enth)g(International)84 1381 y(Workshop)f(on)g (Distribute)n(d)g(A)o(rti\014cial)h(Intel)r(ligenc)o(e)i(\(D)o(AI\))p Fe(,)14 b(pp.)22 b(305{319,)c(Seattle,)d(W)l(ashington,)84 1441 y(1994.)-2 1542 y(17.)24 b(A)d(New)g(Order)h(P)o(arameter)e(for)h (3SA)l(T.)h Fb(Pr)n(o)n(c)n(e)n(e)n(dings)f(of)h(the)h(Twelfth)g (National)g(Confer)n(enc)n(e)g(on)84 1603 y(A)o(rti\014cial)h(Intel)r (ligenc)o(e)h(\(AAAI\))e(Workshop)f(on)h(Exp)n(erimental)g(Evaluation)h (of)f(R)n(e)n(asoning)f(and)84 1663 y(Se)n(ar)n(ch)17 b(Metho)n(ds)p Fe(,)f(pp.)21 b(57{63,)d(Seattle,)d(W)l(ashington,)i (1994.)-2 1764 y(18.)24 b(Utilit)o(y-Based)17 b(T)l(ermination)g(of)h (An)o(ytime)e(Algorithms.)26 b Fb(Pr)n(o)n(c)n(e)n(e)n(dings)18 b(of)i(the)g(Eur)n(op)n(e)n(an)e(Confer-)84 1825 y(enc)n(e)d(on)e(A)o (rti\014cial)h(Intel)r(ligen)q(c)n(e)i(\(ECAI\))d(Workshop)g(on)g(De)n (cision)h(The)n(ory)e(for)h(D)o(AI)f(Applic)n(ations)p Fe(,)84 1885 y(pp.)22 b(88{99,)c(Amsterdam,)13 b(The)j(Netherlands,)f (1994.)-2 1987 y(19.)24 b(An)19 b(Implem)o(en)n(tation)e(of)i(the)f (Con)o(tract)h(Net)f(Proto)q(col)i(Based)e(on)i(Marginal)e(Cost)i (Calculations.)84 2047 y Fb(Pr)n(o)n(c)n(e)n(e)n(dings)i(of)h(the)g (Eleventh)i(National)f(Confer)n(enc)n(e)g(on)f(A)o(rti\014cial)g(Intel) r(ligen)q(c)n(e)j(\(AAAI\))p Fe(,)c(pp.)84 2107 y(256{262,)d(W)l (ashington)e(DC,)g(July)l(,)e(1993.)23 b(\(Acceptance)14 b(rate)j(24\045\))-2 2209 y(20.)24 b(An)19 b(Implem)o(en)n(tation)e(of) i(the)f(Con)o(tract)h(Net)f(Proto)q(col)i(Based)e(on)i(Marginal)e(Cost) i(Calculations.)84 2269 y Fb(Pr)n(o)n(c)n(e)n(e)n(dings)h(of)g(the)h (Twelfth)h(International)g(Workshop)d(on)i(Distribute)n(d)g(A)o (rti\014cial)g(Intel)r(ligenc)o(e)84 2329 y(\(D)o(AI\))p Fe(,)16 b(pp.)21 b(295{308,)e(Hidden)c(V)l(alley)l(,)f(P)o(ennsylv)m (ania,)i(1993.)-2 2431 y(21.)24 b(Automatic)19 b(Co)q(op)q(eration)j (of)e(Area-Distributed)f(Dispatc)o(h)h(Cen)o(ters)g(in)f(V)l(ehicle)f (Routing.)33 b Fb(Pr)n(o-)84 2491 y(c)n(e)n(e)n(dings)20 b(of)f(the)h(International)g(Confer)n(enc)n(e)g(on)g(A)o(rti\014cial)g (Intel)r(ligenc)n(e)i(Applic)n(ations)e(in)g(T)l(r)n(ans-)84 2551 y(p)n(ortation)d(Engine)n(ering)p Fe(,)h(pp.)j(449{467,)e(San)e (Buena)o(v)o(en)o(tura,)d(California,)i(1992.)p eop %%Page: 20 20 20 19 bop -2 150 a Fe(22.)24 b(Automatic)18 b(Co)q(op)q(eration)k(of)e (F)l(actorially)f(Distributed)g(Dispatc)o(h)h(Cen)o(ters)f(in)g(V)l (ehicle)f(Routing.)84 210 y Fb(A)o(bstr)n(act)e(Col)r(le)n(ction)h(of)e (the)h(Joint)f(International)i(Confer)n(enc)n(e)e(on)h(Op)n(er)n (ational)f(R)n(ese)n(ar)n(ch)f(/)h(Man-)84 270 y(agement)k(Scienc)n(e)h (\(EUR)o(O)d(/)g(TIMS\))p Fe(,)f(Helsinki,)e(Finland,)h(1992.)-2 372 y(23.)24 b(A)15 b(Bargaining)h(Net)o(w)o(ork)e(for)i(In)o(telligen) o(t)c(Agen)o(ts.)21 b Fb(Pr)n(o)n(c)n(e)n(e)n(dings)15 b(of)i(the)g(Finnish)g(A)o(rti\014cial)g(Intel)r(li-)84 432 y(genc)n(e)k(Confer)n(enc)n(e)e(\(ST)l(eP\),)h(New)f(Dir)n(e)n (ctions)f(in)h(A)o(rti\014cial)h(Intel)r(ligenc)o(e)p Fe(,)g(V)l(ol.)25 b(3,)18 b(pp.)25 b(173{181,)84 492 y(Esp)q(o)q(o,)18 b(Finland,)e(June,)f(1992.)-2 594 y(24.)24 b(In)o(tro)q(duction)17 b(to)f(the)h(Logistics)f(W)l(orkshop.)23 b Fb(Pr)n(o)n(c)n(e)n(e)n(dings)16 b(of)i(the)g(Finnish)g(A)o (rti\014cial)g(Intel)r(ligenc)o(e)84 654 y(Confer)n(enc)n(e)h(\(ST)l (eP\),)f(New)g(Dir)n(e)n(ctions)f(in)h(A)o(rti\014cial)g(Intel)r (ligenc)o(e)p Fe(,)g(V)l(ol.)j(3,)16 b(pp.)22 b(147-148,)c(Esp)q(o)q (o,)84 714 y(Finland,)e(June,)g(1992.)-2 816 y(25.)24 b(A)13 b(Strategy)g(for)g(Decreasing)g(the)g(T)l(otal)g(T)l(ransp)q (ortation)i(Costs)f(Among)e(Area-Distributed)g(T)l(rans-)84 876 y(p)q(ortation)g(Cen)o(ters.)19 b Fb(Pr)n(o)n(c)n(e)n(e)n(dings)12 b(of)g(the)h(\\Nor)n(dic)f(Op)n(er)n(ations)g(A)o(nalysis)g(in)h(Co)n (op)n(er)n(ation)e(\(NO)o(AS\):)84 936 y(OR)18 b(in)g(Business")g (Confer)n(enc)n(e)p Fe(,)e(T)l(urku,)g(Finland,)g(1991.)-38 1109 y Fc(b.)25 b(In)n(vited)18 b(lectures)f(\(excluding)g(conferences) h(and)h(w)n(orkshops\))-2 1226 y Fe(26.)24 b(Negotiation)15 b(among)g(Computationally)f(Limited)e(Self-In)o(terested)h(Agen)o(ts.) 21 b Fb(University)16 b(of)g(Michi-)84 1286 y(gan)p Fe(.)30 b(Business)19 b(Sc)o(ho)q(ol,)g(Computer)f(and)h(Information)f(Systems) f(Departmen)o(t.)28 b(F)l(unded)18 b(in)o(vited)84 1346 y(talk.)j(3/11/1997)q(.)-2 1448 y(27.)j(Negotiation)17 b(among)f(Computationally)f(Limited)f(Self-In)o(terested)h(Agen)o(ts.) 20 b Fb(University)f(of)e(Penn-)84 1508 y(sylvania)p Fe(.)37 b(The)21 b(Wharton)h(Sc)o(ho)q(ol,)g(Op)q(erations)f(and)h (Information)e(Managemen)o(t)g(Departmen)o(t.)84 1568 y(F)l(unded)d(in)o(vited)d(talk.)21 b(2/26/1997.)-2 1670 y(28.)j(Negotiation)g(among)f(Computationally)f(Limited)f(Self-In)o (terested)h(Agen)o(ts.)41 b Fb(Lund)24 b(University)p Fe(,)84 1730 y(Sw)o(eden.)d(Departmen)o(t)15 b(of)h(Computer)g (Science.)j(F)l(unded)d(in)o(vited)f(talk.)21 b(2/11/1997.)-2 1832 y(29.)j(Negotiation)17 b(among)f(Computationally)f(Limited)f (Self-In)o(terested)h(Agen)o(ts.)20 b Fb(University)f(of)e(Penn-)84 1892 y(sylvania)p Fe(.)j(Computer)10 b(and)i(Information)e(Science)f (Departmen)o(t.)18 b(F)l(unded)10 b(in)o(vited)g(talk.)19 b(11/25/1996.)-2 1994 y(30.)24 b(Negotiation)c(among)g(Computationally) f(Limited)f(Self-In)o(terested)g(Agen)o(ts.)32 b Fb(University)22 b(of)e(Wa-)84 2054 y(terlo)n(o)p Fe(.)i(Departmen)o(t)14 b(of)j(Computer)e(Science.)20 b(F)l(unded)c(job)h(talk.)j(4/11/1996)q (.)-2 2156 y(31.)k(Negotiation)18 b(among)f(Computationally)f(Limited)f (Self-In)o(terested)h(Agen)o(ts.)24 b Fb(University)19 b(of)f(Pitts-)84 2216 y(bur)n(gh)p Fe(.)k(Departmen)o(t)14 b(of)j(Computer)e(Science.)20 b(F)l(unded)c(job)h(talk.)j(4/2/1996)q(.) -2 2317 y(32.)k(Negotiation)12 b(among)f(Computationally)g(Limited)f (Self-In)o(terested)f(Agen)o(ts.)20 b Fb(Columbia)13 b(University)p Fe(.)84 2378 y(Departmen)o(t)i(of)i(Computer)e(Science.) 20 b(F)l(unded)c(job)g(talk.)21 b(3/20/1996)q(.)-2 2479 y(33.)j(Negotiation)15 b(among)g(Computationally)f(Limited)e(Self-In)o (terested)h(Agen)o(ts.)20 b Fb(Washington)d(Univer-)84 2540 y(sity)p Fe(.)k(Departmen)o(t)15 b(of)i(Computer)e(Science.)20 b(F)l(unded)c(job)g(talk.)21 b(3/7/1996.)p eop %%Page: 21 21 21 20 bop -2 150 a Fe(34.)24 b(Automatic)15 b(Co)q(op)q(eration)k(of)d (Area-Distributed)g(Dispatc)o(h)h(Cen)o(ters)f(in)g(V)l(ehicle)e (Routing.)22 b Fb(A)o(t)c(the)84 210 y(Or)n(ganization)13 b(for)f(Ec)n(onomic)h(Co)n(op)n(er)n(ation)e(and)i(Development)i (\(OECD\))d(Scienti\014c)j(Exp)n(ert)e(Gr)n(oup)84 270 y(TT6)24 b(me)n(eting)i(on)e(A)n(dvanc)n(e)n(d)g(L)n(o)n(gistics)g(and) g(Information)g(T)l(e)n(chnolo)n(gy)h(in)f(F)l(r)n(eight)g(T)l(r)n (ansp)n(ort)p Fe(,)84 330 y(W)l(ashington,)18 b(D.C.)d(F)l(unded)h(in)o (vited)f(talk.)21 b(F)l(ebruary)16 b(1994.)-2 432 y(35.)24 b(Computerizing)14 b(Negotiations)h(among)g(F)l(reigh)o(t)f(Carrier)h (Companies.)20 b Fb(A)o(t)d(the)g(Information)f(T)l(e)n(ch-)84 492 y(nolo)n(gy)h(and)g(L)n(o)n(gistics)f(seminar)p Fe(.)21 b(Helsinki)13 b(Univ)o(ersit)o(y)g(of)i(T)l(ec)o(hnology)l(,)g(Esp)q(o) q(o,)i(Finland.)k(F)l(unded)84 552 y(in)o(vited)15 b(talk.)21 b(1992.)-38 725 y Fc(c.)k(Collo)r(quia)17 b(and)i(seminars)-2 842 y Fe(36.)24 b(On)17 b(the)f(Gains)h(and)g(Losses)h(of)f(Sp)q (eculation)f(in)g(Equilibrium)e(Mark)o(ets.)21 b(W)l(ashington)d(Univ)o (ersit)o(y)l(,)84 902 y(Departmen)o(t)d(of)i(Computer)e(Science)g(F)l (acult)o(y)g(Researc)o(h)g(Review)g(presen)o(tation,)h(10/17/1997)q(.) -2 1004 y(37.)24 b(Algorithms)18 b(for)i(Computer)f(Chess.)32 b(W)l(ashington)21 b(Univ)o(ersit)o(y)l(,)c(Departmen)o(t)h(of)i (Computer)f(Sci-)84 1064 y(ence,)c(in)o(vited)g(talk)h(in)g(CS313A)h (Arti\014cial)e(In)o(telligence)e(Lab)q(oratory)l(,)18 b(10/8/1997.)-2 1166 y(38.)24 b(Negotiation)d(among)g(self-in)o (terested)e(agen)o(ts.)35 b(W)l(ashington)22 b(Univ)o(ersit)o(y)l(,)d (Departmen)o(t)g(of)i(Com-)84 1226 y(puter)c(Science)d(F)l(acult)o(y)h (Researc)o(h)h(Review)f(presen)o(tation,)g(10/18/19)q(96.)-2 1327 y(39.)24 b(Presen)o(tation)16 b(at)g(ARP)l(A)f(site)h(visit.)k (Univ)o(ersit)o(y)13 b(of)j(Massac)o(h)o(usetts)g(at)g(Amherst,)e (Departmen)o(t)g(of)84 1388 y(Computer)i(Science,)e(1994.)-2 1489 y(40.)24 b(3)18 b(presen)o(tations)f(to)g(funding)g(sources,)g (including)f(T)l(ec)o(hnology)h(Dev)o(elopmen)o(t)d(Cen)o(ter)i(of)h (Finland)84 1550 y(and)j(industry)g(represen)o(tativ)o(es.)28 b(A)o(t)19 b(the)g(T)l(ec)o(hnical)f(Researc)o(h)h(Cen)o(tre)f(of)i (Finland,)f(Lab)q(oratory)84 1610 y(for)e(Information)e(Pro)q(cessing,) i(1992.)-38 1782 y Fc(11.)25 b(P)n(aten)n(ts)19 b(p)r(ending)p -38 1799 496 2 v 84 1923 a(Metho)r(d)14 b(for)h(Pro)n(v)m(ably)f (E\016cien)n(t)g(Negotiation)f(via)h(Lev)n(eled)f(Commitm)o(e)o(n)n(t)f (Con-)84 1983 y(tracts.)21 b Fe(Submitted)15 b(7/1997.)23 b(P)o(aid)17 b(b)o(y)e(and)i(licensed)e(to)i(BusinessBots,)e(Inc.)84 2124 y Fc(Metho)r(d)30 b(for)g(Optimal)e(An)n(ytime)h(T)-5 b(ask)30 b(Allo)r(cation)f(via)g(P)n(o)n(w)n(erful)j(Con)n(tract)84 2184 y(T)n(yp)r(es.)21 b Fe(Submitted)14 b(7/1997.)24 b(P)o(aid)16 b(b)o(y)g(and)h(licensed)e(to)h(BusinessBots,)g(Inc.)p eop %%Page: 22 22 22 21 bop -38 150 a Fc(12.)25 b(Researc)n(h)18 b(supp)r(ort)p -38 166 525 2 v -38 249 a Fe(1991{)390 b Fc(Sev)n(eral)18 b(signi\014can)n(t)h(sc)n(holarships)477 b Fe(In)o(ternationally)472 309 y(See)16 b(section)g Fb(A)o(war)n(ds)h(and)g(honors)-38 411 y Fe(11/1/199)q(7{1)q(0/)q(31/)q(200)q(0)25 b Fc(An)30 b(op)r(en)f(comp)r(onen)n(t-based)f(arc)n(hitecture)g(for)472 471 y(In)n(ternet)18 b(commerce)1712 411 y Fe(P)o(alo)e(Alto,)g(CA)472 512 y Fb(NIST,)h(US)h(Dep)n(artment)f(of)f(Commer)n(c)n(e)g(A)n(dvanc)n (e)n(d)h(T)l(e)n(chnolo)n(gy)h(Pr)n(o)n(gr)n(am)c(\(A)l(TP\))472 572 y Fe(BusinessBots,)k(Inc.)27 b(\(Dr.)g(Sandholm)17 b(is)h(Chief)g(Scien)o(tist\),)f(Commercene)o(t,)e(CN-)472 633 y(Group,)i(T)l(esserae)f(Information)g(Systems)472 693 y(US$)h(5,000,000)h(and)f(US$)f(5,000,000)i(in)e(matc)o(hing)f (funds)-38 795 y(6/1/1997)q({5/)q(31)q(/20)q(02)74 b Fc(Coalition)34 b(F)-5 b(ormation)34 b(among)g(Self-In)n(terested)472 855 y(Computationally)17 b(Limited)f(Agen)n(ts)1709 795 y Fe(St.)21 b(Louis,)c(MO)472 906 y Fb(NSF)h(CAREER)f(awar)n(d)g (\(Information)g(T)l(e)n(chnolo)n(gy)h(and)f(Or)n(ganizations\))472 966 y Fe(T)l(uomas)g(Sandholm)e(\(PI\))472 1026 y(US$)i(456,000)-38 1128 y(5/1/1997)q({4/)q(30)q(/19)q(99)74 b Fc(Optimal)24 b(Mec)n(hanisms)g(for)h(Negotiation)f(under)472 1188 y(Message)19 b(P)n(assing)h(and)f(Belief)e(Revision)1709 1128 y Fe(St.)k(Louis,)c(MO)472 1239 y Fb(NSF)h(\(Information)f(T)l(e)n (chnolo)n(gy)i(and)e(Or)n(ganizations\))472 1300 y Fe(T)l(uomas)g (Sandholm)e(\(PI\),)h(Ronald)g(Loui)472 1360 y(US$)h(199,000)-38 1461 y(5/1/1997)q({4/)q(30)q(/20)q(00)74 b Fc(High)23 b(P)n(erformance)e(Distributed)g(Ob)s(ject)h(En)n(vi-)472 1522 y(ronmen)n(t)15 b(with)h(Emphasis)f(on)h(Adaptiv)n(e)f(End-to-)472 1582 y(end)k(QoS)g(Guaran)n(tees)1709 1461 y Fe(St.)i(Louis,)c(MO)472 1633 y Fb(D)o(ARP)l(A)h(\(Quorum)f(pr)n(o)n(gr)n(am\))472 1693 y Fe(Guru)c(P)o(arulk)m(ar)g(\(PI\),)e(Douglas)j(Sc)o(hmidt,)c(T)l (uomas)j(Sandholm,)f(Jonathan)i(T)l(urner)472 1753 y(US$)j(650,000)-38 1855 y(8/1/1992)q({12)q(/3)q(1/1)q(992)50 b Fc(Neural)18 b(Nets)g(in)h(Bankruptcy)f(Prediction)472 1915 y(\(YRIAR)-6 b(VI\))1692 1855 y Fe(Esp)q(o)q(o,)18 b(Finland)472 1967 y Fb(T)l(e)n(chnic)n(al)h(R)n(ese)n(ar)n(ch)d(Centr)n(e)i(of)f(Finland) 472 2027 y(L)n(ab)n(or)n(atory)e(for)i(Information)g(Pr)n(o)n(c)n (essing)472 2087 y Fe(Gran)o(tor:)22 b(Remote)15 b(Area)h(Dev)o (elopmen)o(t)d(F)l(und)472 2147 y(T)l(uomas)k(Sandholm)e(\(PI\))472 2207 y(100,000)k(FIM)c(\(US$)h(1)h(v)m(aries)f(b)q(et)o(w)o(een)f(4)i (FIM)e(and)i(6)g(FIM\))472 2267 y(F)l(unding)g(appro)o(v)o(ed)f(but)g (pro)s(ject)g(nev)o(er)f(initiated)472 2328 y(\(PI)i(w)o(en)o(t)e(to)i (graduate)g(sc)o(ho)q(ol\))p eop %%Page: 23 23 23 22 bop -38 150 a Fe(1/31/199)q(2{1)q(2/)q(31/)q(199)q(2)25 b Fc(In)n(telligen)n(t)18 b(Agen)n(ts)h(\(INTELA)n(GENT\))340 b Fe(Esp)q(o)q(o,)18 b(Finland)472 210 y Fb(T)l(e)n(chnic)n(al)h(R)n (ese)n(ar)n(ch)d(Centr)n(e)i(of)f(Finland)472 270 y(L)n(ab)n(or)n (atory)e(for)i(Information)g(Pr)n(o)n(c)n(essing)472 330 y Fe(Sepp)q(o)g(Linnainmaa)f(\(PI\),)g(T)l(uomas)g(Sandholm,)f(and) i(Aarno)g(Leh)o(tola.)472 390 y(500,000)i(FIM)c(\(US$)h(1)h(v)m(aries)f (b)q(et)o(w)o(een)f(4)i(FIM)e(and)i(6)g(FIM\))-38 489 y Fc(13.)25 b(Languages)p -38 505 351 2 v -38 588 a Fe(English,)16 b(Finnish,)g(Sw)o(edish,)f(German)-38 719 y Fc(14.)25 b(Researc)n(h)18 b(summary)p -38 735 562 2 v -38 853 a Fe(Researc)o(h)e(in)o(terests)84 977 y(Arti\014cial)10 b(in)o(telligence,)e(distributed)i(AI,)g(m)o(ultiagen)o(t)e(systems,)j (mac)o(hine)d(learning,)j(resource-)84 1037 y(b)q(ounded)27 b(reasoning,)i(constrain)o(t)d(satisfaction,)i(sc)o(heduling,)f(game)f (theory)l(,)h(distributed)84 1098 y(problem)22 b(solving,)j (negotiation,)g(coalition)e(formation,)g(electronic)f(commerce)o(,)g (com)o(bina-)84 1158 y(torial)15 b(optimization,)d(net)o(w)o(ork)i (pricing,)g(in)o(telligen)o(t)e(real-time)g(systems,)h(normativ)o(e)g (system)84 1218 y(design,)j(industrial)g(soft)o(w)o(are)g(systems)-38 1352 y(Narrativ)o(e)g(researc)o(h)f(summary)84 1440 y(My)22 b(researc)o(h)g(spans)h(a)g(n)o(um)o(b)q(er)d(of)j(sub\014elds)f (within)g(arti\014cial)f(in)o(telligence.)37 b(The)22 b(main)84 1501 y(areas)16 b(are)g(distributed)f(arti\014cial)f(in)o (telligence|esp)q(eci)o(ally)e(m)o(ultiagen)o(t)h(systems|and)h(ma-)84 1561 y(c)o(hine)i(learning,)g(but)h(I)f(also)h(w)o(ork)g(on)g (resource-b)q(ounded)h(reasoning)f(and)g(constrain)o(t)g(satis-)84 1621 y(faction.)k(My)14 b(approac)o(h)g(incorp)q(orates)h(b)q(oth)g (theoretical)e(and)i(rigorous)g(empirical)d(metho)q(ds.)84 1681 y(I)i(ha)o(v)o(e)f(also)i(built)e(large-scale)h(programs)g (including)f(sev)o(eral)g(\014elded)h(industrial)f(AI)g(systems.)84 1741 y(Application)k(domains)f(include)g(distributed)h(planning)g(and)h (sc)o(heduling)f(in)g(man)o(ufacturing)84 1801 y(and)g(airp)q(ort)f (settings,)g(op)q(erativ)o(e)f(truc)o(k)g(and)h(train)g(routing)g(and)g (sc)o(heduling,)f(la)o(w,)g(medical)84 1862 y(diagnosis,)i(and)g (bankruptcy)f(prediction.)k(The)d(con)o(tributions)f(are)g(summarized)d (b)q(elo)o(w.)84 1988 y Fc(Distributed)18 b(Arti\014cial)f(In)n (telligence|Multiagen)n(t)f(Systems.)46 b Fe(Researc)o(h)16 b(in)f(m)o(ul-)84 2048 y(tiagen)o(t)f(systems)e(concerns)i (computational)f(agen)o(ts)h(that)g(act)g(on)g(b)q(ehalf)f(of)h(real)g (agen)o(ts)g(suc)o(h)84 2108 y(as)20 b(companies)e(or)i(individuals.)29 b(By)19 b(co)q(ordinating)h(their)e(actions)i(these)f(agen)o(ts)h(can)f (reac)o(h)84 2168 y(more)13 b(desirable)g(outcomes)g(than)i(they)e (could)h(b)o(y)g(op)q(erating)g(individually)l(.)19 b(The)14 b(imp)q(ortance)84 2229 y(of)j(suc)o(h)e(automated)h(negotiation)g (systems)f(for)h(self-in)o(terested)e(agen)o(ts)i(is)g(lik)o(ely)d(to)j (increase)84 2289 y(as)i(a)e(result)g(of)h(three)f(dev)o(elopmen)o(ts.) j(One)d(is)h(the)f(gro)o(wth)h(of)g(a)g(standardized)f(comm)o(unica-)84 2349 y(tion)i(infrastructure|In)o(ternet,)e(WWW,)h(EDI,)g(NI)q(I,)f(K)o (QML,)h(T)l(elescript,)f(Ja)o(v)m(a)i(etc|o)o(v)o(er)84 2409 y(whic)o(h)c(separately)g(designed)g(agen)o(ts)g(b)q(elonging)h (to)g(di\013eren)o(t)e(organizations)i(can)g(in)o(teract)e(in)84 2469 y(an)20 b(op)q(en)h(en)o(vironmen)o(t)16 b(and)k(safely)f(carry)h (out)g(transactions.)32 b(The)19 b(second)h(is)f(the)g(adv)o(en)o(t)84 2530 y(of)i(small)e(transaction)i(commerce)c(on)k(the)f(In)o(ternet)f (for)i(purc)o(hasing)g(go)q(o)q(ds,)i(information,)84 2590 y(comm)o(unication)c(bandwidth,)k(computation,)f(and)g(storage.)38 b(The)21 b(third)h(is)f(an)h(industrial)84 2650 y(trend)16 b(to)o(w)o(ard)h(virtual)e(en)o(terprises:)20 b(dynamic)15 b(alliances)g(of)h(small,)f(agile)g(en)o(terprises)g(whic)o(h)p eop %%Page: 24 24 24 23 bop 84 150 a Fe(together)19 b(can)g(tak)o(e)g(adv)m(an)o(tage)h (of)f(economies)e(of)i(scale)g(when)g(a)o(v)m(ailable)f(\(e.g.)28 b(resp)q(ond)20 b(to)84 210 y(more)13 b(div)o(erse)f(orders)i(than)g (individual)f(agen)o(ts)h(can\),)g(but)f(do)i(not)f(su\013er)g(from)e (diseconomies)84 270 y(of)17 b(scale.)84 351 y(Multiagen)o(t)c(tec)o (hnology)g(facilitates)f(the)h(automated)g(formation)g(of)h(suc)o(h)f (dynamic)e(alliances)84 411 y(on)j(a)g(p)q(er)g(order)f(basis.)21 b(This)14 b(automation)f(can)g(sa)o(v)o(e)g(lab)q(or)h(time)e(of)h(h)o (uman)g(negotiators,)h(but)84 471 y(in)i(addition,)f(other)h(sa)o (vings)g(are)g(p)q(ossible)g(b)q(ecause)f(computational)g(agen)o(ts)h (are)g(often)g(more)84 532 y(e\013ectiv)o(e)f(at)i(\014nding)h(b)q (ene\014cial)d(short-term)h(coalitions)h(than)g(h)o(umans)f(are)h(in)g (strategically)84 592 y(and)g(com)o(binatorially)d(complex)g(settings.) 84 673 y(In)20 b(co)q(op)q(erativ)o(e)h(distributed)e(problem)g (solving,)i(the)f(system)e(designer)i(imp)q(oses)g(an)g(in)o(ter-)84 733 y(action)i(proto)q(col)g(and)f(a)h(lo)q(cal)f(strategy)h(for)f(eac) o(h)g(agen)o(t.)36 b(The)21 b(approac)o(h)h(is)f(descriptiv)o(e:)84 793 y(with)16 b(the)f(giv)o(en)g(proto)q(col,)h(if)f(the)h(agen)o(ts)g (use)f(the)h(imp)q(osed)f(strategies,)g(then)g(certain)g(so)q(cial)84 853 y(outcomes)f(follo)o(w.)21 b(On)14 b(the)h(other)g(hand,)g(in)f(m)o (ultiagen)o(t)f(systems,)g(the)h(agen)o(ts)i(are)e(pro)o(vided)84 914 y(with)j(an)h(in)o(teraction)e(proto)q(col,)h(but)g(eac)o(h)g(agen) o(t)g(will)f(c)o(ho)q(ose)h(its)g(o)o(wn)g(strategy)l(.)24 b(Th)o(us)17 b(the)84 974 y(proto)q(cols)d(need)e(to)h(b)q(e)f (designed)g Fb(normatively)h Fe(so)g(that)g(the)f(desired)g(lo)q(cal)g (strategies)h(are)f(b)q(est)84 1034 y(for)i(the)g(agen)o(ts|and)g(th)o (us)g(the)f(agen)o(ts)h(will)e(use)i(them|and)e(then)i(certain)f(so)q (cial)g(outcomes)84 1094 y(follo)o(w.)25 b(This)18 b(allo)o(ws)f(the)g (metho)q(ds)g(to)h(function)f(as)h(desired)f(ev)o(en)f(among)i(agen)o (ts)g(that)g(are)84 1154 y(constructed)f(b)o(y)e(separate)i(designers)f (and/or)i(represen)o(t)d(di\013eren)o(t)g(real)h(w)o(orld)g(parties.)84 1235 y(The)g(cen)o(tral)f(theme)e(of)j(m)o(y)e(w)o(ork)i(is)f (negotiation)h(among)g(self-in)o(terested)e(agen)o(ts)i(whose)g(ra-)84 1295 y(tionalit)o(y)c(is)g(limited)d(b)o(y)j(computational)g(complexit) o(y)-5 b(.)17 b(The)12 b(metho)q(ds)g(that)h(I)f(ha)o(v)o(e)g(dev)o (elop)q(ed)84 1356 y(target)k(inheren)o(tly)e(distributed)g(com)o (binatorial)g(problems)g(in)o(v)o(olving)g(agen)o(ts)h(that)h(ha)o(v)o (e)f(dif-)84 1416 y(feren)o(t)f(goals)h(and)g(that)g(attempt)e(to)h (maximize)d(their)i(o)o(wn)i(go)q(o)q(d)h(without)f(concern)f(of)g (global)84 1476 y(w)o(elfare.)20 b(The)12 b(metho)q(ds)h(are)f(needed)g (in)h(settings)g(where)f(the)h(agen)o(ts)g(are)g(self-in)o(terested,)e (and)84 1536 y(there)17 b(is)f(an)h(underlying)f(in)o(tractable)g(com)o (binatorial)f(problem)g(that)i(limits)e(the)h(agen)o(ts')h(ra-)84 1596 y(tionalit)o(y)g(b)q(ecause)g(the)g(problem)f(cannot)i(b)q(e)g (solv)o(ed)e(optimally)g(in)h(practice.)23 b(Applications)84 1657 y(with)f(these)f(t)o(w)o(o)g(c)o(haracteristics)f(include)g (distributed)h(v)o(ehicle)e(routing)i(among)h(indep)q(en-)84 1717 y(den)o(t)16 b(dispatc)o(h)f(cen)o(ters,)g(man)o(ufacturing)f (planning)i(and)h(sc)o(heduling)e(among)g(m)o(ultiple)e(agile)84 1777 y(en)o(terprises)22 b(in)g(sub)q(con)o(tracting)h(net)o(w)o(orks,) g(meeting)e(sc)o(heduling,)i(sc)o(heduling)e(of)i(patien)o(t)84 1837 y(treatmen)o(ts)13 b(across)i(hospitals,)g(classro)q(om)f(sc)o (heduling,)f(planning)h(and)h(sc)o(heduling)e(of)i(m)o(ulti-)84 1897 y(con)o(tractor)20 b(soft)o(w)o(are)g(pro)s(jects,)f(digital)g (libraries,)g(m)o(ultiagen)o(t)e(information)h(gathering)i(on)84 1958 y(the)13 b(W)l(orld)g(Wide)f(W)l(eb,)h(and)g(routing)h(and)f (bandwidth)h(allo)q(cation)f(in)f(m)o(ulti-pro)o(vider)e(m)o(ulti-)84 2018 y(consumer)16 b(computer)g(net)o(w)o(orks,)g(to)i(name)d(just)j(a) f(few.)23 b(In)17 b(suc)o(h)g(com)o(binatorial)e(problems,)84 2078 y(computational)21 b(complexit)o(y)d(precludes)i(exhaustiv)o(e)g (en)o(umeration)f(and)j(ev)m(aluation)g(of)f(all)84 2138 y(p)q(ossible)16 b(outcomes,)e(whic)o(h)h(renders)g(the)g(assumption)g (of)h(p)q(erfect)f(rationalit)o(y)f(inapplicable.)84 2198 y(My)j(w)o(ork)f(extends)h(game)e(theory)i(b)o(y)f(relaxing)g (this)h(assumption)g(via)f(a)h(mo)q(del)f(of)h(b)q(ounded)84 2258 y(rationalit)o(y)j(that)h(quan)o(titativ)o(ely)d(c)o(haracterizes) h(eac)o(h)g(agen)o(t's)i(abilit)o(y)e(to)h(solv)o(e)g(com)o(bina-)84 2319 y(torial)h(problems.)32 b(The)20 b(exp)q(erimen)o(tal)d(part)k(of) g(this)f(w)o(ork)g(uses)h(real)e(data)j(from)d(m)o(ultiple)84 2379 y(collab)q(orating)i(companies.)29 b(My)19 b(con)o(tributions)h (to)g(m)o(ultiagen)o(t)d(systems)h(researc)o(h)h(can)g(b)q(e)84 2439 y(roughly)h(classi\014ed)f(under)g(three)f(in)o(terrelated)g (topics:)27 b(coalition)18 b(formation,)h(con)o(tracting,)84 2499 y(and)e(con)o(tract)f(execution.)84 2580 y Fb(Co)n(alition)f (formation.)20 b Fe(I)13 b(dev)o(elop)q(ed)f(a)h(normativ)o(e)e(theory) i(of)g(coalitions)g(among)g(self-in)o(teres-)84 2640 y(ted,)i(computationally)e(limited)f(agen)o(ts.)21 b(It)14 b(states)h(whic)o(h)f(agen)o(ts)h(should)g(form)f(coalitions|)p eop %%Page: 25 25 25 24 bop 84 150 a Fe(i.e.)19 b(co)q(ordinate)12 b(their)f(actions|and) h(whic)o(h)f(coalition)h(structures)f(are)h(stable.)20 b(I)11 b(analytically)84 210 y(pro)o(v)o(ed)h(these)g(prescriptions,)g (whic)o(h)g(are)g(based)h(on)g(the)f(p)q(erformance)f(pro\014les)h(of)h (the)f(agen)o(ts')84 270 y(problem)g(solving)h(algorithms.)19 b(The)13 b(prescriptions)g(di\013er)f(signi\014can)o(tly)g(from)g (those)h(for)h(fully)84 330 y(rational)k(agen)o(ts.)25 b(My)16 b(theory)i(includes)e(a)h(formal)g(application)g(indep)q(enden) o(t)f(domain)g(clas-)84 390 y(si\014cation)i(for)f(b)q(ounded)h (rational)g(agen)o(ts.)24 b(It)17 b(also)h(determines)d(the)h (relationships)h(b)q(et)o(w)o(een)84 451 y(these)h(domain)g(classes)g (and)h(those)f(from)f(t)o(w)o(o)h(theories)g(of)g(fully)f(rational)i (agen)o(ts.)27 b(In)18 b(addi-)84 511 y(tion,)f(I)g(studied)g(a)h(real) e(w)o(orld)h(distributed)g(v)o(ehicle)e(routing)i(problem)f(to)h (determine)e(where)84 571 y(in)20 b(the)f(domain)f(classi\014cation)i (suc)o(h)f(a)h(real)f(problem)e(falls.)31 b(The)19 b(analysis)h(sho)o (w)o(ed)f(ho)o(w)h(a)84 631 y(traditional)d(domain)e(classi\014cation)h (fails)g(to)h(incorp)q(orate)g(this)f(problem.)k(The)c(exp)q(erimen)o (ts)84 691 y(found)k(the)f(b)q(est)h(coalitions)f(among)g(routing)g (agen)o(ts,)h(and)g(determined)d(their)h(stabilit)o(y)l(.)29 b(It)84 752 y(turns)21 b(out)g(that)g(with)f(rational)h(agen)o(ts,)g (all)f(coalitions)h(are)f(b)q(est)h(o\013)g(merging)e(while)h(with)84 812 y(b)q(ounded)e(rational)e(agen)o(ts)h(this)f(is)g(not)g(the)g (case.)21 b(Another)16 b(in)o(teresting)f(\014nding)i(is)f(that)h(m)o (y)84 872 y(prescriptions)e(regarding)h(coalition)f(structures)h(are)f (v)o(ery)g(close)g(to)g(the)h(coalition)f(structures)84 932 y(that)j(h)o(uman)f(negotiators)h(form)f(in)g(the)g(routing)h (problem.)24 b(A)o(t)17 b(the)g(same)f(time)g(they)h(di\013er)84 992 y(drastically)e(from)f(the)g(prescriptions)h(that)h(classic)e(game) g(theory)h(mak)o(es)e(regarding)j(fully)e(ra-)84 1053 y(tional)g(agen)o(ts.)21 b(This)14 b(leads)g(me)e(to)i(b)q(eliev)o(e)e (that)i(m)o(y)e(normativ)o(e,)f(quan)o(titativ)o(e)i(theories)g(also)84 1113 y(ha)o(v)o(e)k(descriptiv)o(e)f(accuracy)h(\(whic)o(h)g(w)o(as)h (not)h(striv)o(ed)d(for)i(as)h(a)f(goal\).)26 b(Finally)l(,)16 b(I)h(formally)84 1173 y(sho)o(w)o(ed)f(a)g(parado)o(x:)21 b(an)16 b(agen)o(t)g(ma)o(y)e(b)q(e)h(b)q(etter)g(o\013)h(with)g(a)g (slo)o(w)o(er)e(algorithm)h(than)h(a)f(faster)84 1233 y(one.)84 1314 y Fb(Contr)n(acting.)37 b Fe(In)21 b(the)g(con)o(tract)g (net)g(framew)o(ork,)f(agen)o(ts)h(solv)o(e)g(problems)f(b)o(y)g (iterativ)o(ely)84 1374 y(making)e(decisions)g(ab)q(out)i(issues)f(suc) o(h)f(as)i(the)e(allo)q(cation)h(of)g(tasks)h(among)e(themselv)o(es.)26 b(I)84 1434 y(extended)19 b(the)g(con)o(tract)h(net)f(for)h(use)g (among)f(self-in)o(terested,)g(computationally)f(b)q(ounded)84 1495 y(agen)o(ts)13 b(in)f(sev)o(eral)f(w)o(a)o(ys.)19 b(First,)12 b(I)g(incorp)q(orated)h(rigorous)g(micro)q(economic)c (principles)h(whic)o(h)84 1555 y(lead)19 b(to)g(b)q(etter)f(global)h (solutions)g(b)o(y)g(facilitating)e(more)h(precise)f(tradeo\013s)j (among)f(agen)o(ts.)84 1615 y(Second,)26 b(I)e(extended)f(the)g(proto)q (col)i(to)f(encompass)g(lev)o(eled)e(commi)o(tm)o(en)o(t)e(con)o (tracts.)45 b(I)84 1675 y(pro)o(v)o(ed)18 b(that)h(these)f(increase)g (the)g(e\016ciency)f(of)h(deals,)h(and)g(sometimes)d(enable)i(deals)g (that)84 1735 y(are)k(imp)q(ossible)d(via)i(traditional)g(full)f (commitm)o(e)o(n)o(t)e(con)o(tracts.)36 b(Recen)o(tly)19 b(I)i(also)h(sho)o(w)o(ed)84 1796 y(the)i(ordering)g(of)g(sev)o(eral)f (di\013eren)o(t)g(lev)o(eled)e(commitm)n(en)o(t)g(con)o(tracting)i (proto)q(cols)i(based)84 1856 y(on)e(ho)o(w)g(m)o(uc)o(h)d(the)j(agen)o (ts)g(lie)e(in)h(equilibrium)d(\(less)j(is)g(b)q(etter\).)40 b(The)22 b(proto)q(cols)i(under)84 1916 y(comparison)16 b(di\013er)g(based)h(on)g(the)f(sync)o(hronous)h(\(sim)o(ultaneous)e (decommitti)o(ng\))f(vs.)22 b(asyn-)84 1976 y(c)o(hronous)g(\(sequen)o (tial)e(decommitti)o(ng\))f(implem)o(en)o(tation.)33 b(I)21 b(also)h(enric)o(hed)e(the)h(proto)q(col)84 2036 y(with)i(coun)o(terprop)q(osals)i(among)e(agen)o(ts)g(and)h(the)f(use)g (of)g(dynamically)e(c)o(hosen)i(stage)g(of)84 2097 y(commitm)o(en)n(t.) 18 b(These)d(allo)o(w)g(for)h(a)f(more)f(e\016cien)o(t)f(searc)o(h)i (of)h(desirable)e(con)o(tracts.)21 b(Third,)15 b(I)84 2157 y(pro)o(v)o(ed)i(that)g(con)o(tracting)g(one)g(task)h(at)f(a)h (time)d(do)q(es)i(not)h(lead)f(to)g(a)h(global)f(optim)o(um,)d(and)84 2217 y(dev)o(elop)q(ed)i(con)o(tract)h(t)o(yp)q(es)f(that)i(remedy)c (this)j(problem:)k(a)c(globally)f(optimal)g(task)h(allo)q(ca-)84 2277 y(tion)g(is)f(guaran)o(teed)h(in)f(a)h(\014nite)f(n)o(um)o(b)q(er) e(of)j(con)o(tracts)g(without)f(bac)o(ktrac)o(king.)21 b(These)c(new)84 2337 y(con)o(tract)g(t)o(yp)q(es)g(include)f(atomic)g (con)o(tracts)h(of)g(m)o(ultiple)d(tasks,)k(sw)o(ap)g(con)o(tracts,)e (and)i(con-)84 2398 y(tracts)g(among)f(more)e(than)j(t)o(w)o(o)f(agen)o (ts.)24 b(I)16 b(also)i(iden)o(ti\014ed)d(sev)o(eral)h(tradeo\013s)i (in)f(an)h(agen)o(t's)84 2458 y(lo)q(cal)g(delib)q(eration)f(and)h (comm)o(uni)o(cation)d(sc)o(heduling,)h(and)i(dev)o(elop)q(ed)f(metho)q (ds)g(for)h(mak-)84 2518 y(ing)d(the)g(asso)q(ciated)g(decisions)f (dynamically)l(.)k(Finally)l(,)13 b(I)i(pro)o(v)o(ed)f(that)h(the)f (classical)g(Vic)o(krey)84 2578 y(auction)19 b(proto)q(col)g(fails)f (to)h(promote)f(truthful)g(bidding)g(and)h(to)g(a)o(v)o(oid)f(coun)o (tersp)q(eculation)84 2638 y(among)f(computationally)f(b)q(ounded)i (agen)o(ts)g(in)e(man)o(y)g(settings)h(where)g(it)f(is)h(often)g(b)q (eliev)o(ed)p eop %%Page: 26 26 26 25 bop 84 150 a Fe(to)19 b(ha)o(v)o(e)e(these)h(desirable)g(c)o (haracteristics.)25 b(William)16 b(Vic)o(krey)l(,)g(the)h(in)o(v)o(en)o (tor)g(of)i(the)e(proto-)84 210 y(col,)k(receiv)o(ed)e(the)h(1996)i (Nob)q(el)e(Prize)g(in)g(economics.)33 b(My)20 b(w)o(ork)g(do)q(es)i (not)f(con\015ict)f(with)84 270 y(his)f(original)f(claims.)26 b(Instead)18 b(m)o(y)f(results)h(sho)o(w)h(that)g(there)f(ha)o(v)o(e)f (b)q(een)i(mistak)o(es)d(in)i(ho)o(w)84 330 y(Vic)o(krey's)f(results)i (ha)o(v)o(e)f(b)q(een)i(in)o(terpreted,)e(generalized,)g(and)i(applied) e(in)h(computational)84 390 y(m)o(ultiagen)o(t)14 b(systems.)84 471 y(Also)20 b(within)g(the)g(\014eld)g(of)g(con)o(tracting,)h(I)f (designed)g(and)g(dev)o(elop)q(ed)g(a)g(distributed)g(asyn-)84 532 y(c)o(hronous)g(protot)o(yp)q(e)f(system,)e(TRA)o(CONET,)h(for)h (exploring)f(automated)g(task)h(allo)q(cation)84 592 y(in)e(v)o(ehicle)d(routing.)22 b(I)16 b(solv)o(ed)g(sev)o(eral)f (previously)h(unaddressed)h(implem)o(en)n(tation)d(issues.)22 b(I)84 652 y(sho)o(w)o(ed)17 b(that)g(requiring)e(replies)g(is)i(not)g (feasible)e(among)h(self-in)o(terested)f(agen)o(ts.)22 b(Rather,)16 b(a)84 712 y(proto)q(col)h(based)f(on)g(time-outs)f(is)g (necessary)l(.)21 b(F)l(urthermore,)14 b(I)h(reduced)g(the)g(task)h (announc-)84 772 y(ing)c(problem)f(to)h(a)h(w)o(ell)d(kno)o(wn)i (dilemma)d(from)i(game)g(theory)l(,)h(the)g(tragedy)g(of)g(the)g (commons.)84 833 y(It)20 b(follo)o(ws)g(that)h(agen)o(ts)g(are)f(prone) h(to)f(saturation)i(b)o(y)d(incoming)g(messages.)33 b(I)20 b(dev)o(elop)q(ed)84 893 y(sev)o(eral)d(metho)q(ds)g(for)g(a)o(v)o (oiding)g(this)h(problem.)23 b(I)17 b(also)h(dev)o(elop)q(ed)f(optimal) f(algorithms)g(for)84 953 y(terminating)f(the)h(con)o(tracting)g(pro)q (cess.)84 1034 y Fb(Contr)n(act)e(exe)n(cution.)22 b Fe(I)12 b(dev)o(elop)q(ed)g(a)g(metho)q(d)g(that)h(enables)f(agen)o(ts) g(to)h(carry)f(out)h(exc)o(hanges)84 1094 y(without)18 b(enforcemen)o(t.)j(It)16 b(is)h(based)h(on)f(splitting)g(the)g(exc)o (hange)f(in)o(to)h(parts.)24 b(I)17 b(pro)o(v)o(ed)f(the)84 1154 y(equilibrium)c(strategies)i(that)h(complete)e(the)h(exc)o(hange)g (in)g(the)h(minim)o(al)d(n)o(um)o(b)q(er)h(of)i(steps|)84 1214 y(for)g(coun)o(table)f(and)h(uncoun)o(table)g(go)q(o)q(ds,)h(and)f (for)g(isolated)f(and)i(non-isolated)f(exc)o(hanges.)20 b(I)84 1275 y(analyzed)15 b(the)f(role)g(of)h(real)g(time)d(in)j(con)o (tract)f(execution)g(and)h(the)f(usefulness)h(of)g(unenforced)84 1335 y(deadlines)k(as)h(w)o(ell)e(as)i(lateness)f(p)q(enalt)o(y)g(sc)o (hedules.)29 b(I)19 b(also)h(devised)e(a)i(non)o(trivial,)e(sound,)84 1395 y(and)d(complete)d(quadratic)i(algorithm)f(for)h(sequencing)g (indep)q(enden)o(t)f(deliv)o(eries,)f(and)i(pro)o(v)o(ed)84 1455 y(that)e(in)o(terdep)q(enden)o(t)d(go)q(o)q(ds)k(cannot)f(b)q(e)f (sequenced)f(in)h(p)q(olynomial)e(time.)18 b(Finally)l(,)10 b(I)h(sho)o(w)o(ed)84 1515 y(that)17 b(the)f(new)g(unenforced)g(exc)o (hange)g(metho)q(d)f(a)o(v)o(oids)h(unfair)h(renegotiation.)84 1596 y(In)11 b(addition)h(to)f(m)o(y)e(researc)o(h)i(on)g(self-in)o (terested)f(agen)o(ts,)i(I)e(ha)o(v)o(e)h(also)h(w)o(ork)o(ed)e(on)i (co)q(op)q(erativ)o(e)84 1657 y(negotiation.)21 b(Sp)q(eci\014cally)l (,)11 b(I)h(participated)g(in)g(designing)g(and)h(dev)o(eloping)f(a)h (large-scale)f(pro-)84 1717 y(tot)o(yp)q(e)k(system,)f(DIS-DSS,)h(for)h (co)q(op)q(erativ)o(e)f(distributed)f(sc)o(heduling)g(of)i(airp)q(ort)g (resources)84 1777 y(and)g(generalized)e(man)o(ufacturing)g(job-shops.) 84 1858 y Fc(F)-5 b(uture)13 b(Directions.)18 b Fe(Distributed)10 b(systems)g(consisting)h(of)g(m)o(ultiple)d(self-in)o(terested)h(agen)o (ts)84 1918 y(will)18 b(b)q(ecome)f(increasingly)g(common,)g(and)i(the) f(appropriate)h(use)g(of)g(limited)c(computation)84 1978 y(and)k(comm)o(unication)c(resources)j(in)g(suc)o(h)g(systems)f(will)g (b)q(e)i(of)f(ma)s(jor)g(imp)q(ortance.)26 b(I)17 b(ha)o(v)o(e)84 2038 y(adv)m(anced)j(the)e(state)h(of)g(the)f(art)h(in)f(m)o(ultiagen)o (t)e(systems,)i(and)h(plan)g(to)g(con)o(tin)o(ue)e(w)o(ork)i(in)84 2099 y(this)i(gro)o(wing)h(\014eld.)35 b(In)20 b(the)h(short)h(term,)e (I)g(w)o(ould)h(lik)o(e)f(to)h(dev)o(elop)f(coalition)g(formation)84 2159 y(proto)q(cols)f(that)f(dynamically)d(guide)i(the)h(agen)o(ts)g (to)g(the)f(optimal)f(solutions)i(prescrib)q(ed)f(b)o(y)84 2219 y(m)o(y)11 b(theory)l(,)i(and)g(to)g(extend)f(the)h(theory)f(to)h (encompass)g(probabilistic)e(p)q(erformance)h(pro\014les.)84 2279 y(There)20 b(are)h(a)f(host)i(of)e(other)h(in)o(teresting)e(op)q (en)i(questions)f(in)g(coalition)g(formation,)g(and)h(I)84 2339 y(ha)o(v)o(e)16 b(receiv)o(ed)d(the)j(NSF)f(CAREER)h(a)o(w)o(ard)g (to)g(further)g(study)g(coalition)f(formation)h(among)84 2400 y(self-in)o(terested)h(agen)o(ts.)29 b(Within)18 b(con)o(tracting,)g(I)g(plan)h(to)g(study)g(the)f(e\013ects)g(of)h(lo)q (ok)m(ahead)84 2460 y(regarding)g(the)f(con)o(tracting)f(pro)q(cess)i (and)g(external)e(ev)o(en)o(ts.)25 b(There)17 b(are)i(also)f(sev)o (eral)f(op)q(en)84 2520 y(issues)h(regarding)f(the)g(lev)o(eled)e (commitm)o(e)o(n)o(t)f(con)o(tracting)j(proto)q(cols.)26 b(In)17 b(addition,)g(I)g(w)o(an)o(t)84 2580 y(to)c(explore)f(b)q (elief)g(revision)f(in)i(m)o(ultiagen)o(t)d(systems|a)h(recen)o(t)h (pro)s(ject)g(for)h(whic)o(h)f(I)g(receiv)o(ed)84 2640 y(another)17 b(NSF)f(gran)o(t)h(\(Ronald)f(Loui)h(is)f(a)h(co-PI)f(on)h (this\).)p eop %%Page: 27 27 27 26 bop 84 150 a Fe(I)18 b(w)o(ould)f(also)h(lik)o(e)e(to)i(dev)o (elop)e(a)i(real)f(w)o(orld)g(m)o(ulti-en)o(terprise)d(man)o (ufacturing)i(sc)o(heduling)84 210 y(system)f(based)h(on)g(the)f (normativ)o(e)f(principles)g(I)h(ha)o(v)o(e)g(dev)o(elop)q(ed.)20 b(Curren)o(tly)14 b(suc)o(h)i(systems)84 270 y(are)i(b)q(oth)h (ine\016cien)o(t)c(and)j(manipulable.)24 b(Other)17 b(application)g (areas)i(of)e(particular)h(in)o(terest)84 330 y(to)13 b(me)e(include)g(In)o(ternet)g(commerc)o(e,)f(electricit)o(y)f(mark)o (ets,)i(net)o(w)o(ork)g(pricing,)h(and)h(micro)q(eco-)84 390 y(nomic)i(principles)f(in)i(net)o(w)o(ork)f(services)g(suc)o(h)g (as)i(information)e(gathering,)h(digital)f(libraries,)84 451 y(and)i(the)f(allo)q(cation)h(of)f(computation)g(and)h(storage.)84 577 y Fc(Mac)n(hine)e(Learning.)48 b Fe(Within)11 b(mac)o(hine)g (learning,)h(I)g(ha)o(v)o(e)g(completed)f(three)g(main)h(pieces)84 637 y(of)20 b(researc)o(h.)30 b(First,)20 b(I)f(empiricall)o(y)d (studied)k(m)o(ultiagen)o(t)d(reinforcemen)o(t)f(learning.)30 b(I)19 b(used)84 697 y(agen)o(ts)g(with)g(lo)q(okup)g(tables)f(and)h (recurren)o(t)f(neural)g(net)o(w)o(ork)g(memorie)o(s,)e(di\013eren)o(t) i(length)84 757 y(histories,)24 b(and)g(di\013eren)o(t)d(exploration)i (sc)o(hedules.)40 b(I)22 b(iden)o(ti\014ed)g(the)g(m)o(ultiple)e (reasons|)84 818 y(i.e.)j(t)o(yp)q(es)17 b(of)h(in)o(terpla)o(y)d(of)j (the)f(agen)o(ts')g(learning,)g(exploration)g(strategies,)g(and)g(sto)q (c)o(hastic)84 878 y(action)d(selectors|wh)o(y)f(classical)g (reinforcemen)o(t)d(learning)k(algorithms)e(are)i(not)g(guaran)o(teed) 84 938 y(to)j(con)o(v)o(erge)e(in)h(suc)o(h)g(settings.)84 1018 y(Second,)e(I)e(join)o(tly)g(refuted)g(a)i(pro)q(of)g(in)f (decision)f(tree)g(induction.)20 b(It)12 b(claimed)f(that)i(trees)g (with)84 1079 y(few)o(er)18 b(lea)o(v)o(es)e(are)i(lik)o(ely)e(to)i(ha) o(v)o(e)g(higher)g(generalization)f(accuracy)h(than)g(trees)g(with)g (more)84 1139 y(lea)o(v)o(es.)25 b(W)l(e)17 b(sho)o(w)o(ed)h(that)g (suc)o(h)g(claims)e(are)i(imp)q(ossible)e(without)i(assumptions)g(ab)q (out)h(the)84 1199 y(prior)f(distribution)e(of)i(target)f(concepts.)24 b(This)18 b(also)g(casts)f(doubts)h(on)g(m)o(ultiple)c(attempted)84 1259 y(pro)q(ofs)k(of)f(Occam's)d(Razor,)i(y)o(et)f(it)h(do)q(es)h(not) g(dispro)o(v)o(e)e(the)h(principle)f(itself.)84 1340 y(Third,)i(I)f(ha)o(v)o(e|in)f(collab)q(oration)j(with)e(medical)e(do)q (ctors|compared)j(the)f(generalization)84 1400 y(accuracy)i(of)g(sev)o (eral)f(inductiv)o(e)f(learning)h(metho)q(ds)h(in)f(a)h(real)g(medical) d(diagnosis)k(domain.)84 1460 y(W)l(e)h(analyzed)f(the)g(follo)o(wing)h (m)o(ultiv)m(ariate)d(metho)q(ds:)27 b(linear)19 b(regression,)h (logistic)f(regres-)84 1520 y(sion,)j(3-la)o(y)o(er)d(feedforw)o(ard)i (neural)f(net)o(w)o(orks)g(with)g(bac)o(kpropagation,)i(decision)e (trees,)h(k-)84 1580 y(nearest)h(neigh)o(b)q(or)f(classi\014ers,)g(and) h(a)f(comp)q(osite)g(system)e(called)h(MCS.)h(W)l(e)g(also)g(exp)q (eri-)84 1641 y(men)o(ted)f(with)h(feature)g(pruning.)38 b(The)21 b(results)h(con)o(v)o(ey)o(ed)d(a)j(selectiv)o(e)d(sup)q (eriorit)o(y)i(of)h(the)84 1701 y(metho)q(ds,)17 b(and)g(a)h(strong)g (correlation)f(of)g(the)g(classi\014cation)g(accuracies)f(of)i(the)e (metho)q(ds)h(in)84 1761 y(di\013eren)o(t)f(problems.)84 1841 y(In)e(addition)f(to)h(these)f(three)g(main)f(e\013orts,)j(I)e(ha) o(v)o(e)g(w)o(ork)o(ed)f(on)i(applying)g(genetic)e(algorithms)84 1902 y(to)17 b(a)g(design)f(problem,)e(and)j(on)g(applying)f(neural)g (net)o(w)o(orks)g(to)g(bankruptcy)h(prediction.)84 1982 y Fc(F)-5 b(uture)17 b(Directions.)i Fe(Learning)c(is)f(an)h(imp)q (ortan)o(t)e(capabilit)o(y)g(in)h(m)o(ultiagen)o(t)d(systems)i(but)84 2042 y(there)i(is)f(a)h(lac)o(k)f(of)h(normativ)o(e)e(con)o(v)o(ergen)o (t)g(algorithms)h(for)h(learning)g(among)f(self-in)o(terested)84 2102 y(agen)o(ts.)22 b(As)15 b(I)g(ha)o(v)o(e)g(sho)o(wn,)h(learning)f (in)g(suc)o(h)g(settings)h(in)o(v)o(olv)o(es)d(man)o(y)h(complicated)g (issues)84 2163 y(in)20 b(addition)h(to)f(those)h(of)g(single)e(agen)o (t)i(learning.)33 b(I)20 b(plan)g(to)h(con)o(tin)o(ue)e(m)o(y)f(w)o (ork)i(on)h(w)o(ell)84 2223 y(founded)13 b(algorithms)f(for)h(m)o (ultiagen)o(t)d(learning.)20 b(I)12 b(w)o(ould)g(also)i(lik)o(e)c(to)j (con)o(tin)o(ue)f(m)o(y)e(researc)o(h)84 2283 y(on)18 b(the)f(theory)f(of)i(induction,)e(and)h(to)h(apply)e(mac)o(hine)f (learning)i(tec)o(hniques)e(in)i(large-scale)84 2343 y(systems)e(and)i(real)f(w)o(orld)g(problems,)f(esp)q(ecially)f(in)i (medical)e(and)j(\014nancial)f(domains.)84 2469 y Fc(Resource-Bounded)j (Reasoning.)48 b Fe(In)17 b(addition)g(to)h(the)f(m)o(ultiagen)o(t)d (researc)o(h)j(on)h(lim-)84 2530 y(ited)24 b(rationalit)o(y)g(outlined) f(ab)q(o)o(v)o(e,)j(I)e(ha)o(v)o(e)g(w)o(ork)o(ed)g(on)g(single)g(agen) o(t)h(resource-b)q(ounded)84 2590 y(reasoning.)31 b(I)19 b(dev)o(elop)q(ed)f(an)i(optimal)e(metho)q(d)g(for)i(terminating)d (incomplete)g(decision)h(al-)84 2650 y(gorithms)d(\(con)o(v)o(erted)e (in)o(to)i(an)o(ytime)e(algorithms\))h(whose)i(p)q(erformance)d (pro\014les)i(are)g(condi-)p eop %%Page: 28 28 28 27 bop 84 150 a Fe(tioned)11 b(on)h(execution)d(so)j(far.)20 b(The)10 b(conditioning)h(problem)f(is)g(not)i(tractable)e(for)h (optimization)84 210 y(problems.)84 291 y Fc(F)-5 b(uture)18 b(Directions.)i Fe(I)15 b(plan)h(to)g(con)o(tin)o(ue)e(researc)o(h)h (on)h(normativ)o(e)e(metho)q(ds)h(for)h(con)o(trol-)84 351 y(ling)e(reasoning,)g(and)g(new)g(mo)q(dels)e(of)i(limited)d (rationalit)o(y)l(.)19 b(These)14 b(ha)o(v)o(e)e(direct)h(implications) 84 411 y(for)k(lo)q(cal)f(delib)q(eration)g(con)o(trol)g(and)h(proto)q (col)g(design)f(in)g(m)o(ultiagen)o(t)e(systems.)84 538 y Fc(Constrain)n(t)h(Satisfaction.)48 b Fe(In)11 b(3-satis\014abilit)o (y)l(,)h(the)g(clauses)g(to)g(v)m(ariables)g(ratio)g(has)g(clas-)84 598 y(sically)h(b)q(een)h(used)g(as)h(an)f(order)g(parameter.)20 b(Through)15 b(rigorous)g(statistical)e(testing)h(I)g(found)84 659 y(a)20 b(second)g(order)f(parameter.)30 b(I)19 b(used)g(a)h(com)o (bination)e(of)i(the)f(t)o(w)o(o)g(parameters)g(to)h(predict)84 719 y(satis\014abilit)o(y)13 b(and)i(the)f(run)o(times)e(of)i(t)o(w)o (o)g(satis\014abilit)o(y)f(testing)h(algorithms)f(from)g(the)h(litera-) 84 779 y(ture.)21 b(The)15 b(accuracy)f(w)o(as)i(signi\014can)o(tly)d (higher)i(than)h(with)e(either)g(parameter)g(alone.)21 b(I)14 b(also)84 839 y(used)h(the)f(parameters)f(to)i(condition)f(the)h (p)q(erformance)e(pro\014le)h(of)h(an)f(an)o(ytime)f(satis\014abilit)o (y)84 899 y(testing)k(algorithm)e(that)h(I)g(devised.)84 980 y Fc(F)-5 b(uture)26 b(Directions.)36 b Fe(I)21 b(w)o(ould)h(lik)o (e)e(to)i(use)g(order)g(parameters)f(in)g(v)m(ariable)g(and)i(v)m(alue) 84 1041 y(ordering)17 b(heuristics)e(in)h(constrain)o(t)g(satisfaction) h(algorithms.)k(I)16 b(plan)g(to)h(dev)o(elop)e(a)i(generic)84 1101 y(constrain)o(t)d(satisfaction)f(engine)g(that)h(statistically)e (manages)h(the)g(algorithm's)f(p)q(erformance)84 1161 y(pro\014les.)22 b(This)16 b(engine)f(w)o(ould)h(also)g(b)q(e)g(useful) g(as)g(a)g(lo)q(cal)g(reasoning)h(mo)q(dule)d(in)i(m)o(ultiagen)o(t)84 1221 y(systems,)h(for)h(example)d(in)j(the)f(m)o(ulti-en)o(terprise)d (man)o(ufacturing)j(sc)o(heduling)g(system)f(that)84 1281 y(w)o(as)h(suggested)g(as)g(future)f(w)o(ork)g(ab)q(o)o(v)o(e.)84 1408 y Fc(Industrial)27 b(In)n(telligen)n(t)e(Systems.)47 b Fe(In)23 b(addition)g(to)h(the)f(protot)o(yp)q(e)h(systems)e(that)i (I)84 1469 y(built)17 b(for)g(researc)o(h)f(purp)q(oses,)i(I)e(ha)o(v)o (e)g(w)o(ork)o(ed)h(at)g(applied)f(researc)o(h)g(cen)o(ters)g(in)h(dev) o(eloping)84 1529 y(three)f(large-scale)g(industrial)g(AI)f(systems.)84 1610 y(First,)f(I)g(co-dev)o(elop)q(ed)g(a)h(truc)o(k)e(routing)i(and)g (sc)o(heduling)e(application,)h(EPO,)g(for)h(a)g(forestry)84 1670 y(compan)o(y)l(.)k(I)13 b(designed)h(and)g(implem)o(en)o(te)o(d)d (the)i(mo)q(dule)f(that)i(in)o(telligen)o(tly)d(allo)q(cates)i(w)o (eekly)84 1730 y(transp)q(ortation)18 b(tasks)e(among)f(da)o(ys.)21 b(I)15 b(also)h(dev)o(elop)q(ed)f(half)g(of)h(the)f(mo)q(dule)f(that)i (allo)q(cates)84 1790 y(distributed)k(w)o(o)q(o)q(d)i(batc)o(hes)e (among)g(factories)h(based)f(on)h(m)o(ultiple)c(criteria.)33 b(The)20 b(system)84 1850 y(w)o(as)g(c)o(hosen)f(one)g(of)g(the)g(y)o (ear's)f(17)i(most)f(inno)o(v)m(ativ)o(e)e(AI)i(applications)g(in)o (ternationally)e(at)84 1911 y(the)d(1995)h(Conference)e(on)i(Inno)o(v)m (ativ)o(e)d(Applications)h(of)h(AI.)e(It)i(has)g(b)q(een)g(\014elded)f (and)h(its)g(use)84 1971 y(has)j(paid)g(bac)o(k)f(its)g(dev)o(elopmen)o (t)d(costs.)84 2052 y(Next)i(I)g(built)f(a)i(train)g(routing)f(and)h (sc)o(heduling)f(application)g(for)h(the)f(same)f(compan)o(y)l(.)20 b(That)84 2112 y(system)15 b(is)g(curren)o(tly)f(b)q(eing)i(\014elded.) k(Both)c(the)f(truc)o(k)g(and)h(train)f(transp)q(ortation)j(problems)84 2172 y(w)o(ere)i(com)o(binatorial)f(and)i(the)f(instances)h(v)o(ery)e (large.)34 b(They)21 b(w)o(ere)e(c)o(haracterized)g(b)o(y)h(n)o(u-)84 2232 y(merous)14 b(hard)g(to)h(enco)q(de)f(constrain)o(ts)g(from)g (domain)f(exp)q(erts.)20 b(The)14 b(solutions)h(used)f(domain)84 2293 y(sp)q(eci\014c)i(algorithms.)84 2373 y(Third,)h(I)f(participated) g(in)g(the)h(dev)o(elopmen)o(t)c(of)k(a)g(rule-based)g(p)q(ension)g(la) o(w)g(exp)q(ert)f(system)84 2434 y(using)f(the)f(commerc)o(ial)d (XIPlus)i(shell,)g(and)i(in)f(the)g(dev)o(elopmen)o(t)d(and)j(testing)g (of)h(a)f(compiler)84 2494 y(for)k(transforming)f(XIPlus)g(rules)g(in)o (to)g(C-co)q(de.)26 b(The)17 b(exp)q(ert)g(system)f(is)i(in)f(daily)g (use)g(at)h(an)84 2554 y(insurance)e(compan)o(y)l(.)p eop %%Page: 29 29 29 28 bop 84 150 a Fc(F)-5 b(uture)17 b(Directions.)i Fe(I)13 b(plan)h(to)h(use)f(m)o(y)e(industrial)h(exp)q(erience)g(to)h (fo)q(cus)h(m)o(y)d(researc)o(h)h(to-)84 210 y(w)o(ard)g(problems)d(of) i(practical)f(imp)q(ortance.)19 b(I)11 b(w)o(ould)h(also)g(lik)o(e)e (to)i(use)g(some)f(of)h(m)o(y)d(normativ)o(e)84 270 y(metho)q(ds)16 b(in)g(real)g(w)o(orld)g(systems.)84 397 y Fc(Summary)-5 b(.)46 b Fe(I)18 b(ha)o(v)o(e)g(pro)q(duced)g(concrete)f(results)h(in)g (sev)o(eral)f(areas)i(of)f(computer)f(science)84 457 y(and)h(w)o(ork)o(ed)e(in)g(a)h(v)m(ariet)o(y)f(of)h(application)f (domains.)22 b(My)16 b(researc)o(h)g(metho)q(dology)h(includes)84 518 y(b)q(oth)j(theoretical)d(and)i(empirical)c(w)o(ork)j(as)h(w)o(ell) e(as)i(application)f(system)f(building.)26 b(I)18 b(ha)o(v)o(e)84 578 y(published)e(extensiv)o(ely)d(in)j(highly)f(selectiv)o(e)f (forums,)h(giv)o(en)g(n)o(umerous)g(in)o(vited)f(talks,)i(b)q(een)84 638 y(in)o(vited)f(to)i(consult)f(on)h(in)o(teresting)e(AI)g(problems,) f(and)j(giv)o(en)f(tutorials)g(at)h(leading)f(confer-)84 698 y(ences.)21 b(I)16 b(ha)o(v)o(e)f(also)i(successfully)e(\014elded)h (m)o(ultiple)d(AI)i(systems.)p eop %%Trailer end userdict /end-hook known{end-hook}if %%EOF