Solving multiple-objective flexible job shop problems by evolution and local search


Finding realistic schedules for flexible job shop problems has attracted many researchers recently due to its nondeterministic polynomial time (NP) hardness. In this paper, we present an efficient approach for solving the multiple-objective flexible job shop by combining evolutionary algorithm and guided local search (GLS). Instead of applying random local search to find neighboring solutions, we introduce a GLS procedure to accelerate the process of convergence to Pareto-optimal solutions. The main improvement of this combination is to help diversify the population toward the Pareto front. A branch and bound algorithm for finding the lower bounds of multiple-objective solutions is also proposed. Experimental results indicate that the multiple-objective Pareto-optimal solutions of our algorithms dominate previous designs for solving the same benchmarks while incurring less computational time.