报告人:马继平

时间：2016年3月25日（周五）下午3:00—5:00

地点：西安交大文管大楼874教室

报告题目：Recursive Decision Theory

内容摘要：In decision theory , subjective probability theory change the foundation of statistics and decision theory. Fentti and Savage made excellent contributions. Following their studies recently Gilbow, Schemidler introduce the CBR(case-based reasoning) in decision theory and axiomatize it. The CBDT(case-based decision theory) provide a deep insight into human decision-making process. Therefore they think the case-based theory is the less problematic knowledge theory, because there are many unsolved theoretical problems in rule-based decision theory, inconsistency, induction of justification... . In fact there is a internal paradox in their studies. The core part of CDBT is a similarity function, this will lead the theoretical paradox.

Based the studies and our viewpoint, we try to develop a frame to model the rule-generating process. As Savage pointing out, human’s objectives are foundation of statistics. Theories and rules are derived from the objectives through hierarchically abstracting data from the interactions with the world.

This paper is divided into the following parts: 1. Review, mainly focus on the case-based decision theory. 2.We assume that all the plans, actions are recursive divided into atom-level elements. 3.First-step abstraction, according to human utility(inner comfort) category the interaction data. The data is direct about human body, in time and space dimensions. First-step abstraction path is the shortest and most reliable. 4. Multi-step abstraction is a recursive process based on the first-step abstraction.

报告人简介：马继平，西安交通大学金禾中心应用经济学博士生。

欢迎其他校院感兴趣的师生参加！

更多活动请访问 http://jinhe.xjtu.edu.cn/