Title :
Solving MAX-SAT problems using a memetic evolutionary meta-heuristic
Author :
Boughaci, Dalila ; Drias, Habiba ; Benhamou, Belaid
Author_Institution :
Univ. of Sci. & Technol., Algiers, Algeria
Abstract :
Genetic algorithms are a population-based meta-heuristic. They have been successfully applied to many optimization problems. However, premature convergence is an inherent characteristic of such classical genetic algorithms that makes them incapable of searching numerous solutions of the problem domain. A memetic algorithm is an extension of the traditional genetic algorithm. It uses a hill climbing search technique to reduce the likelihood of the premature convergence. In this paper, a memetic approach is studied for the NP-hard satisfiability problems, in particular for its optimization version namely MAX-SAT. Our evolutionary approach applies a search technique to further improve the fitness of individuals in the genetic population. Basically, the approach combines local search heuristics with crossover operators. The method is tested and various experimental results show that memetic algorithm performs better than the classical genetic algorithms for most benchmark problems.
Keywords :
computability; computational complexity; genetic algorithms; heuristic programming; problem solving; search problems; MAX-SAT problem solving; NP-hard satisfiability problem; genetic algorithm; hill climbing search technique; memetic algorithm; memetic evolutionary meta-heuristic; premature convergence; Artificial intelligence; Benchmark testing; Computational complexity; Evolutionary computation; Genetic algorithms; Heuristic algorithms; Large scale integration; Performance evaluation; Simulated annealing; Variable speed drives;
Conference_Titel :
Cybernetics and Intelligent Systems, 2004 IEEE Conference on
Print_ISBN :
0-7803-8643-4
DOI :
10.1109/ICCIS.2004.1460462