Designing decision-making agents for electronic markets

随着电子商务在当今社会中的作用越来越大，智能计算机系统(代理)正被要求在购买、销售和交易商品、服务和信息的电子市场中代表用户做出决定。这样的决策需要详细了解用户的喜好；因此，高效而准确的方法来了解用户的偏好并利用不完全信息进行推理是这项工作的关键组成部分。例如，在代理与其他代理协商或竞争的情况下(代表其他用户)，其他关注包括设计有效的谈判策略和模拟对手的偏好和目标的能力。这个研究计划的总体目标是开发改进的技术，以便智能代理在电子市场中代表用户做出决策。特别强调的是，这种代理需要使他们的行为适应每个用户的偏好和目标。预计每年至少有3名研究生参与这个研究计划，这项工作有望对电子商务、偏好启发、决策理论推理、用户建模和自动谈判等领域的研究做出贡献。

With the increased role of electronic commerce in today's society, intelligent computer systems (agents) are being called upon to make decisions on a user's behalf in electronic markets where goods, services and information are purchased, sold and traded. Such decision making requires detailed knowledge of the user's preferences; therefore, efficient and accurate methods for learning about the user's preferences and for reasoning with incomplete information are critical components of this work.  In the case of settings where an agent is negotiating with or competing against other agents (acting on the behalf of other users), additional concerns include the design of effective negotiation strategies and the ability to model the opponent's preferences and goals.The overall goal of this research program is to develop improved techniques for intelligent agents to use in making decisions on behalf of users in electronic markets.  Particular emphasis is placed on the need for such agents to adapt their behaviour to the preferences and goals of each individual user.  It is anticipated that at least 3 graduate students per year will be involved in this research program and the work is expected to make contributions to research in the areas of electronic commerce, preference elicitation, decision-theoretic reasoning, user modeling and automated negotiation.