Constraint Satisfaction Problems: CSP Formalisms and Techniques

Dr. Khaled Ghedira Nov 1st Thursday, 12:30 to 1:30pm 209 Comp. Sci. Bldg. Abstract This talk is a summary of the book being edited by Wiley. The concept of ‘constraint’ is very old. It was used in many domains to represent various problems: from school examples such as the n-queens problem to industrial applications such as scheduling problems. The objective of this talk is to introduce how to express such problems as a set of constraints involving a set of variables, each one of them having its own domain of values. For this, we propose to present a simple and generic formalism, the Constraint Satisfaction Problem (CSP). As this formalism is not sufficient to represent all kinds of these problems, different extensions will be detailed such as maximal CSP, Distributed CSP, Dynamic CSP, Constraint Satisfaction and Optimization Problem (CSOP). The objective of our talk is also to introduce how to solve such problems by using two families of methods: backtrack based algorithms such as Forward checking and metaheuristics such as tabu search and evolutionary algorithms . This talk will be illustrated by some relevant examples and applications.