A Method for Flexible Job-Shop Scheduling using Genetic Algorithm

This paper focused on solving a flexible job-shop scheduling problem. Because this problem is known as NP-hard, methods using meta-heuristics especially genetic algorithm (GA) have been actively proposed. Although it is possible to obtain solutions of large problems in a reasonable time by those methods, the quality of the solutions decreases as the scale of the problem increases. Hence, taking advantage of knowledge included in heuristic dispatching rules in the optimization by GA was proposed, and its effectiveness was proven. However, in this method, the two kinds of selection required in flexible job-shop production, machine selection and job selection, were carried out sequentially. Because this may result in insufficient search of the solution space, this paper provided a method using GA in which those two selections were performed at once. The method was applied to an example and it was shown that better solutions could be obtained.