Intelligent Agents and Multi-Agent Systems: 5th Pacific Rim International Workshop on Multi-Agents, PRIMA 2002, Tokyo, Japan, August 18-19, 2002. Proceedings (Lecture Notes in Computer Science, 2413)

Autonomous Agents And Multi-agent Systems Are Computational Systems In Which Several (semi-)autonomous Agents Interact With Each Other Or Work Together To Perform Some Set Of Tasks Or Satisfy Some Set Of Goals. These Systems May Involve Computational Agents That Are Homogeneous Or Heterogeneous, They May Involve Activities On The Part Of Agents Having Common Or Distinct Goals, And They May Involve Participation On The Part Of Humans And Intelligent Agents. This Volume Contains Selected Papers From Prima 2002, The 5th Paci?c Rim International Workshop On Multi-agents, Held In Tokyo, Japan, On August 18–19, 2002 In Conjunction With The 7th Paci?c Rim International Conference On Arti?cial Intelligence (pricai-02). Prima Is A Series Of Workshops On - Tonomous Agents And Multi-agent Systems, Integrating Activities In The Asian And Paci?c Rim Countries. Prima 2002 Built On The Great Success Of Its Pre- Cessors, Prima’98 In Singapore, Prima’99 In Kyoto, Japan, Prima 2000 In Melbourne, Australia, And Prima 2001 In Taipei, Taiwan. We Received 35 Submissions To This Workshop From 10 Countries. Each Paper Was Reviewed By Three Internationally Renowned Program Committee Members. After Careful Reviews, 15 Papers Were Selected For This Volume. We Would Like To Thank All The Authors Who Submitted Papers To The Workshop. We Would Also Like To Thank All The Program Committee Members For Their Splendid Work In Reviewing The Papers. Finally, We Thank The Editorial Sta? Of Springer-verlag For Publishing This Volume In The Lecture Notes In Arti?cial Intelligence. Architecture, Models, And Mechanisms -- An Architecture For Normative Reactive Agents -- A Real-time Agent Architecture: Design, Implementation And Evaluation -- Individual Level Analysis Using Decision Making Features In Multiagent Based Simulation -- False-name-proof Multi-unit Auction Protocol Utilizing Greedy Allocation Based On Approximate Evaluation Values -- Coordination, Negotiation, And Organization -- Developing Alternative Mechanisms For Multiagent Coordination -- An Operational Semantics For Negotiating Agents -- An Organizational Metamodel For The Design Of Catalogues Of Communicative Actions -- Principles For Dynamic Multi-agent Organizations -- Agent-based Interface -- Conducting The Disambiguation Dialogues Between Software Agent Sellers And Human Buyers -- Mutual Learning Of Mind Reading Between A Human And A Life-like Agent -- Automatic Short Story Generator Based On Autonomous Agents -- Meta-level Architecture For Executing Multi-agent Scenarios -- Applications -- Application-oriented Flow Control In Agent-based Network Middleware -- Continuous Truck Delivery Scheduling And Execution System With Multiple Agents -- Designing Agents For Context-rich Textual Information Tasks. Kazuhiro Kuwabara, Jaeho Lee (eds.). Includes Bibliographical References And Index. Also Available Via The World Wide Web. Intelligent Agents and Multi-Agent Systems Preface Program Committee Table of Contents An Architecture for Normative Reactive Agents 1 Introduction 2 The Agent Architecture 3 The Planning Algorithm 4 The Replanning Algorithm 5 A Model of Normative Reasoning 6 Reactivity to Norms 7 Related Work and Conclusions References A Real-Time Agent Architecture: Design, Implementation and Evaluation 1 Introduction 2 Logical Architecture 2.1 Perception 2.2 Cognition 2.3 Action 3 Physical Architecture 3.1 Message Passing 3.2 Overall Physical Architecture 3.3 Sensor Workers and the Sensor Administrator 3.4 The Cognition Workers 3.5 The Task Administrator, the Scheduler Worker and Executor Workers 3.6 An Agent-Based Real-Time Arcade Game 4 Task Scheduling in Real-Time Agents 4.1 Theoretical Analysis 4.2 Experimental Results 5 Related Work 6 Concluding Remarks References Individual Level Analysis Using Decision Making Features in Multiagent Based Simulation 1 Introduction 2 Individual Level Analysis 2.1 Action Uncertainty 2.2 Dominant Input Attributes 3 Criterion of Initial Learning Maturity 3.1 ABM of Deregulated Electric Power Market 3.2 Experiments on ABM of Deregulated Electric Power Market 4 Criterion of Model Discrimination 5 Visualization of Attention Tendency in a Group 6 Discussion 7 Conclusion Acknowledgments References False-Name-Proof Multi-unit Auction Protocol Utilizing Greedy Allocation Based on Approximate Evaluation Values 1 Introduction 2 Greedy ALlocation (GAL) Protocol 2.1 Single Step Case 2.2 Multiple Step Case with Decreasing Average Marginal Utility 2.3 General Case 3 Iterative Reducing (IR) Protocol 4 Evaluations 5 Discussions 6 Conclusions References Appendix A: Details of GAL Protocol Appendix B: Outline of Proof (GAL Is False-Name-Proof) Developing Alternative Mechanisms for Multiagent Coordination 1 Introduction 2 Task Structures and the Enables Relationship 3 Coordination Mechanisms for Enabling Task Relationships 3.1 Avoidable Dependency 3.2 Sacrifice Avoidable Dependency 3.3 Coordination by Reservation 3.4 Predecessor Deadline Commitment 4 Implementation 5 Experimental Results 6 Conclusion and Future Work 6.1 Future Work References An Operational Semantics for Negotiating Agents 1 Introduction 2 Formal Model and Notations 2.1 Negotiation Model 2.2 Negotiating Agent Model 2.3 Negotiation Structure 3 Negotiation Language 4 Operational Semantics 4.1 Transition Rules 4.2 MAS Transition Rules 5 Related Works 6 Conclusion Acknowledgment References An Organizational Metamodel for the Design of Catalogues of Communicative Actions 1 Introduction 2 An Organizational Perspective on ACLs 2.1 The Role/Interaction/Communicative Action Model 2.2 The Organizational Model of the FIPA ACL 3 Design Method of the ACL for a Multiagent System 4 Example: The ACL of an Online Stock Broker 5 Discussion References Principles for Dynamic Multi-agent Organizations 1 Introduction 2 Adapting the Architecture of the Organization 2.1 Some Problems with Static Organizations 2.2 How Does Social Organizations Manage These Problems? 2.3 The Three Principles Applied to Multi-agent Systems 3 Experiments 3.1 Magique 3.2 Three Experiments for Three Principles 4 Conclusion References Conducting the Disambiguation Dialogues between Software Agent Sellers and Human Buyers 1 Introduction 2 The Dialogue System Architecture 2.1 The Buyer-Seller Interaction System 2.2 The Agents 2.3 The Dialogue Model 3. Resolution of Referential Ambiguities 3.1 The Knowledge Models for Disambiguation 3.2 The Strategies of Disambiguation 3.2.1 The Guessing Strategy 3.2.2 The Filtering Strategy 3.2.3 The Recommending Strategy 3.2.4 The Asking More Hints Strategy 4 The Implementation and Experiments 4.1 The Implementation 4.2 A Dialogue Scenario 4.3 Experiments 4.3.1 Finding Optimal Parameters for the Disambiguation Strategies 4.3.2 Effects on Dialogue against Different Combinations 5 Conclusion and Future Work Reference Mutual Learning of Mind Reading between a Human and a Life-Like Agent 1 Introduction 2 Learning of Mind Mapping 2.1 What Should Be Learned? 2.2 Learning in an Agent 2.3 Success Rate and Finish Condition of Learning 3 Mutual Mind Reading Game 4 Implementation 5 Experiments 5.1 Observing Mutual Learning 5.2 Improving Human Learning by Strategic Mind Transition 6 Conclusion References Automatic Short Story Generator Based on Autonomous Agents 1 Introduction 2 Related Work 3 Autonomous Characters 3.1 Personalities 3.2 Emotions 3.3 Motivations 4 A Model for Generating Consistent Episodes 4.1 Moods 4.2 Roles 5 A Model for Interaction 6 Design and Implementation 6.1 System Architecture 6.2 Character Agent Architecture 6.3 Application 7 Conclusions and Future Work References Meta-level Architecture for Executing Multi-agent Scenarios 1 Introduction 2 Backgrounds 2.1 Scenario Description Language 2.2 Requirements for Executing Multi-agent Scenarios 3 Meta-level Architecture for Scenario Execution 3.1 Web-Style GUI for Customizing Scenarios 3.2 Meta-level Architecture 3.3 Sequence of Scenario Execution 4 Experiments on Meta-level Architecture by Evacuation Simulation 4.1 Procedure of Experiment and Scenarios for Agents 4.2 Test for Meta-level Architecture 5 Conclusions References Application-Oriented Flow Control in Agent-Based Network Middleware 1 Introduction 2 Architecture of Dynamic Networking 3 Agent Framework for Dynamic Networking 4 Application-Oriented Flow Control Mechanism 4.1 Concept of DFCA 4.2 Design of DFCA 4.3 Implementation 4.4 Experiments 5 Conclusion References Continuous Truck Delivery Scheduling and Execution System with Multiple Agents Introduction 2 Truck Delivery Scheduling Poblem 3 Solving Tuck Dlivery Scheduling Problem in the Mobile Environment 3.1 Basic Approach to the Problem Solving 3.2 A Truck Delivery Scheduling System Using Multiple Agents 3.3 Function Composition and System-Level Algorithm 3.3.1 Generation of the Initial Delivery Plan 3.3.2 Processing Performed by Each Agent 3.4 Dynamic Re-scheduling at the Time of Delivery Execution 4 Experiments, Evaluations and Analysis 4.1 Details of the Proving System 4.2 Evaluations and Discussion 5 Conclusion References Designing Agents for Context-Rich Textual Information Tasks Introduction 2 A Context-Rich Data Domain 2.1 Characterizing the Information Task 2.2 Implications on Agent System Design 3 Information Retrival Task 3.1 Agents for Mediating Information Retrival 3.2 Guiding Users with Concepts and Terms 3.3 Performance Evaluation and Discussions 4 Information Extraction Task 4.1 Domain Knowledge and Task Analysis 4.2 Agents for Extracting Personnel Information 4.3 Experimental Results and Discussions 5 A View on Agent System Designing for Textual Tasks 5.1 Agent Role 5.2 Agent Group Interactions 5.3 Discussions 6 Summary References Author Index Autonomous agents and multi-agent systems are computational systems in which several (semi-)autonomous agents interact with each other or work together to perform some set of tasks or satisfy some set of goals. These systems may involve computational agents that are homogeneous or heterogeneous, they may involve activities on the part of agents having common or distinct goals, and they may involve participation on the part of humans and intelligent agents. This volume contains selected papers from PRIMA 2002, the 5th Paci?c Rim International Workshop on Multi-Agents, held in Tokyo, Japan, on August 18{u2013}19, 2002 in conjunction with the 7th Paci?c Rim International Conference on Arti?cial Intelligence (PRICAI-02). PRIMA is a series of workshops on - tonomous agents and multi-agent systems, integrating activities in the Asian and Paci?c Rim countries. PRIMA 2002 built on the great success of its pre- cessors, PRIMA{u2019}98 in Singapore, PRIMA{u2019}99 in Kyoto, Japan, PRIMA 2000 in Melbourne, Australia, and PRIMA 2001 in Taipei, Taiwan. We received 35 submissions to this workshop from 10 countries. Each paper was reviewed by three internationally renowned program committee members. After careful reviews, 15 papers were selected for this volume. We would like to thank all the authors who submitted papers to the workshop. We would also like to thank all the program committee members for their splendid work in reviewing the papers. Finally, we thank the editorial sta? of Springer-Verlag for publishing this volume in the Lecture Notes in Arti?cial Intelligence This book constitutes the refereed proceedings of the 5th Pacific Rim International Conference on Multi-Agents, PRIMA 2002, held in Tokyo, Japan in August 2002. The 15 revised full papers presented were carefully reviewed and selected from 35 submissions. The book offers topical sections on architectures, models, and mechanisms; coordination, negotiation, and organization; agent-based inter faces; and applications