Solving constraint satisfaction sequencing problems by iterative repair

Many constraint satisfaction problems involve sequencing constraints, where the aim is to nd a sequence for a domain of values such that all the constraints on the sequence are satis ed. Specialised techniques have been developed to tackle this problem within the constraint programming framework using constructive, backtracking search. In this paper we investigate local search techniques to tackle this problem. By taking advantage of the structure of the sequencing problem we show that within a local search framework we can reduce the size of the search space and the number of constraints which are required to be satis ed. We present SwapGenet, an iterative repair-based algorithm designed speci cally for solving the constraint satisfaction sequencing problem. We present results of an empirical evaluation demonstrating the superiority of SwapGenet over Genet on the car sequencing problem.