Computer Science and Decision Theory : Applications of Notions of Consensus Description of the Research Issues in the Project 1 Introduction

Abstract:Many modern computer science problems involve issues that decision theorists have addressed for years, in particular issues involving consensus and associated order relations. Applications of methods of decision theory to problems of computer science place great strain on these methods due to the sheer size of the problems addressed, limitations on information possessed, and sequential nature of repeated applications. Hence, there is great need to develop a new generation of methods to satisfy these requirements of CS applications. In turn, the new methods will provide powerful tools of use in problems of the social sciences (economics, political science, etc.) to which methods of decision theory have traditionally been applied as well as to newer areas of application of decision theory such as in policy-making concerning emerging diseases or bio-terrorism. This project seeks to explore the connections between computer science and decision theory, develop new decision-theory-based methodologies relevant to the scope of modern CS problems, and investigate their applications to problems of computer science and of the social sciences. We shall build the project around notions of consensus that are so central in modern decision theory. The project is a joint effort between DIMACS, the Center for Discrete Mathematics and Theoretical Computer Science, headquartered at Rutgers University, and LAMSADE, the Laboratoire d’Analyse et Modélisation de Systèmes pour l’Aide à la Décision, based at Université Paris IX Dauphine.

Computer Science