Rules of encounter: designing conventions for automated negotiation among computers 论文

1994引用 1007
Multi-Agent Systems and NegotiationLogic, Reasoning, and Knowledge

摘要

Part 1 Machines that make deals: the premise machine encounters social engineering for machines scenarios how does this differ from Al? how does this differ from game theory? Part 2 Interaction mechanisms: the negotiation problem in different domains attributes of negotiation mechanisms assumptions incentive compatibility. Part 3 Task-oriented domains: domain definition attributes and examples a negotiation mechanism evaluation of the negotiation mechanism an alternative, one-step protocol mechanisms that maximize the product of utilities the bottom line. Part 4 Deception-free protocols: non-manipulable negotiation mechanisms probabilistic deals subadditive domains concave domains modular domains summary of incentive compatible mechanisms the bottom line. Part 5 State-oriented domains: side-effects in encounters domain definition attributes and examples a negotiation mechanism worth of a goal conflict resolution semi-co-operative deals in non-conflict situations unified negotiation protocols (UNP) multi-plan deals the hierarchy of deal types - summary unbounded worth of a goal - tidy agents the bottom line. Part 6 Strategic manipulation: negotiation with incomplete information incomplete information about worth of goals using the revelation principle to re-design the mechanisms the bottom line. Part 7 Worth-oriented domains: goal relaxation domain definition one agent best plan negotiation over sub-optimal states examples of worth functions the bottom line. Appendices: strict/tolerant mechanisms some related work proofs.

相关技术

暂无数据

相关事件

暂无数据

相关文章

暂无数据