Multiagent decision techniques for electronic commerce and environmental management problems

本研究计划的主要目的是发展多智能体决策方法，并将其应用于电子商务和环境管理问题。代理可以是人类决策者，也可以是自主软件代理。特别是，正式的多主体决策技术将被调查和计算机系统支持这些技术将被开发。提出的研究计划由三个部分组成：进一步发展冲突解决方法的图模型；在诸如全流域水资源管理系统和电信网络等各个领域产生的共同资源管理问题中，设计公平和有效地在相互竞争的用途和利益攸关方之间分配有限资源的方法；并研究电子市场（e-marketplaces）在新经济中供应链中的作用，以及电子市场的两个特定方面：在线拍卖和代理之间的谈判。所研究的决策技术将应用于电子商务、环境管理和其他领域的实际决策问题。在本研究中开发的技术和工具可用于帮助决策者解决多主体决策问题，并可被从业人员、研究人员和决策者用于设计电子市场，提高电子市场中公司的谈判绩效，以及在竞争用途和利益相关者之间公平有效地分配稀缺资源。研究生和本科生将接受培训，研究开发多智能体决策技术，设计和实施决策支持系统，并将新的和改进的技术应用于具有挑战性的工程问题。

The principal aim of the proposed research program is to develop multiagent decision making methodologies with application to electronic commerce and environmental management problems. Agents can be human decision makers or autonomous software agents. In particular, formal multiagent decision techniques will be investigated and computerized systems to support these techniques will be developed. The proposed research program is composed of three components: further development of the graph model for conflict resolution methodology; design of methods for allocating fairly and efficiently limited resources among competing uses and stakeholders in common-pool resource management problems arising in various areas, such as basin-wide water resources management systems and telecommunication networks; and study of roles of electronic marketplaces (e-marketplaces) in supply chains in the new economy and two particular aspects of e-marketplaces: online auction and negotiations among agents. The decision making techniques studied will be applied to practical decision problems in electronic commerce, environmental management, and other areas. The techniques and tools developed in the proposed research can be used to assist decision makers who are involved in multiagent decision problems and can be utilized by practitioners, researchers, and decision makers for designing e-marketplaces, improving negotiation performance of firms in e-marketplaces, and allocating scarce resources fairly and efficiently among competing uses and stakeholders. Graduate and undergraduate students will be trained for conducting research on developing multiagent decision techniques, designing and implementing decision support systems, and applying the new and improved techniques to challenging engineering problems.