Insect Swarm Algorithms for Dynamic MAX-SAT Problems

Abstract

The satisfiability (SAT) problem and the maximum satisfiability problem (MAX-SAT) were among the first problems proven to be \(\mathcal{N P}\)-complete. While only a limited number of theoretical and real-world problems come as instances of SAT or MAX-SAT, many combinatorial problems can be encoded into them. This puts the study of MAX-SAT and the development of adequate algorithms to address it in an important position in the field of computer science. Among the most frequently used optimization methods for the MAX-SAT problem are variations of the greedy hill climbing algorithm. This chapter studies the application to dynamic MAX-SAT (i.e. MAX-SAT problems with structures that change over time) of the swarm based metaheuristics ant colony optimization and wasp swarm optimization algorithms, which are based in the real life behavior of ants and wasps, respectively. The algorithms are applied to several sets of static and dynamic MAX-SAT instances and are shown to outperform the greedy hill climbing and simulated annealing algorithms used as benchmarks.