Genetic algorithm versus scatter search and solving hard MAX-W-SAT problems

Abstract

The genetic algorithm approach is well mastered nowadays. It has been applied to a diverse array of classical and real world problems such as the time-tabling and the traveling salesman problems. The scatter search, yet another evolutionary approach has been developed recently. Some known versions have been tested on problems such as the quadratic assignment and the linear ordering problems. In this paper, both approaches are studied for the NP-hard satisfiability problem namely SAT and its optimization version MAX-W-SAT. Empirical tests are performed on Johnson benchmarks and the numerical results show a slight performance difference in favor of the scatter search. © Springer-Verlag Berlin Heidelberg 2001. © Springer-Verlag Berlin Heidelberg 2001.