用于复杂多智能体系统资源分配的协商理论的研究

Automated negotiation is an important research area bridging together artificial intelligence, economics, and game theory. It has received prominent attention in recent years and its importance is widely acknowledged since intelligent agents that negotiate with each other on behalf of human users are expected to lead to more efficient negotiations. However, existing research mainly focuses on automated negotiation in relatively simple environment. The current lack of design of negotiation agents for complex environment constitutes a significant impediment to applying automated negotiation in realistic domains, such as dynamic electronic commerce and large scale distributed systems. This project aims to develop fundamentally new negotiation theory for resource allocation in complex multiagent systems. Specially, this project will 1) develop algorithms for computing equilibrium strategies in complex negotiation scenarios, which will provide insights and theoretical foundations for developing negotiation agents; 2) design and implement next generation heuristic strategies for negotiation agents in realistic environments with incomplete information; and 3) propose negotiation theory for agents that can proficiently negotiate with people. This proposed research is unique and different from existing research and it will open new directions of both theory and application of automated negotiation.We plan to publish over 10 papers at top international journals and conferences. We expect that the research results will be highly respected by experts in this filed.

本项目将研究用于复杂多智能体系统资源分配的协商理论和关键技术。通过研究，本项目将建立面向复杂环境的协商均衡策略理论，弥补为各种复杂应用环境下协商智能体设计提供理论基础和指导；发展应用于复杂环境的启发式协商策略决策理论，为现实复杂环境下协商智能体的设计提供有效方法；设计用于与人协商的智能体协商决策策略，为自动协商的发展提供新的方向。最终，该研究将弥补当前研究对复杂环境协商理论研究的不足，.推动自动协商在更复杂环境中（比如电子商务、大规模分布式系统、以及复杂人机交互环境）的应用，为智能体协商理论的发展和应用提供新的途径。项目预期在在国际重要学术期刊以及会议发表论文10篇以上，围绕该项目的研究成果得到国际同行的认可。

本项目研究了用于复杂多智能体系统资源分配的协商理论和关键技术。研究内容包括：1) 面向复杂环境的协商均衡策略理论；2) 应用于复杂环境的启发式协商策略决策理论；3) 用于与人协商的智能体协商策略设计。.本项目从2013年1月开始，至2015年12月结束，由中国科学院计算技术研究所安波课题组执行。课题组在项目的执行期间严格遵照项目计划书的要求，并充分发挥研究人员的创新作用和科研探索精神，取得了预期成果，现已经超过计划书中的所有预期目标。项目研究成果共发表学术论文13篇，其中中国计算机学会推荐期刊会议目录中A类5篇，B类8篇。通过本项目的研究，能够为智能体协商理论的发展和应用提供了新的途径。