Extremal optimization for solving job shop scheduling problem

Job shop is one of a well known NP-hard optimization problems. In this paper, extremal optimization is proposed for job shop scheduling. Extremal optimization is an evolutionary meta-heuristic method that consecutively substitutes undesirable variables in current solution with a random value and evolves itself toward optimal solution. For EO, the quality of generated initial solution plays an important role in convergence rate and reaching global optimum; hence GT method is utilized for initial solution. This algorithm is implemented on several sample problems on LA datasets and show that optimal solution can be reached quickly on most of the datasets.