Abstract

Many problems from the \real world" can be eciently expressed as constraint satisfaction problems (CSPs). Most of these can be naturally modelled using n-ary (or non-binary) constraints. Representing problems with n-ary constraints and reasoning with them is therefore very important in constraint satisfaction. However, issues regarding n-ary constraints have been neglected compared to binary constraints. The reasons were the simplicity of dealing with binary constraints compared to nary and the fact that any non-binary CSP can be encoded into an equivalent binary. This thesis makes an empirical and theoretical study on representation and solution methods for n-ary CSPs. The results we present demonstrate the importance of the choice of representation and reasoning techniques in n-ary problems.

Citations