A local search algorithm for a SAT representation of scheduling problems

Abstract

This paper presents the application of a local search algorithm for a logical representation of the Job Shop Scheduling Problem (JSSP). This logical representation represents the JSSP transformed as a satisfiability problem (SAT). The proposed algorithm uses a local search in a wide neighborhood. This algorithm, called Walk Wide Search - SAT, is a variant of the WalkSAT algorithm. This search is possible because the included tabu list prevents an excessive number of repetitions of movements during the search process. This paper describes the algorithm and compares results of Walk Wide Search - SAT to WalkSAT. © Springer-Verlag Berlin Heidelberg 2007.