Enhanced Walksat with Finite Learning Automata For MAX-SAT
Noureddine Bouhmala · Redalyc (Universidad Autónoma del Estado de México) · 2014
"Researchers in artificial intelligence usually adopt the co nstraint satisfaction problem and the Satisfiability paradigms as their preferred methods whe n solving various real worlds deci- sion making problems. Local search algorithms used to tackl e different optimization problems that arise in various fields aim at finding a tactical interpla y between diversification and inten- sification to overcome local optimality while the time consu mption should remain acceptable. The Walksat algorithm for the Maximum Satisfiability Proble m (MAX-SAT) is considered to be the main skeleton underlying almost all local search al gorithms for MAX-SAT. This paper introduces an enhanced variant of Walksat using Finit e Learning Automata. A bench- mark composed of industrial and random instances is used to c ompare the effectiveness of the proposed algorithm against state-of-the-art algorithms."