A genetic algorithm with a machine order-based representation scheme for a class of job shop scheduling problem

Abstract

In this paper, we propose a genetic algorithm (GA) with a machine order-based representation scheme (MORS) and apply it to a class of job shop scheduling problems (JSSP), the n/m/J/C/sub max/ problems, where n/spl ges/3*m. The proposed approach uses a special genotype-to-phenotype decoding method which guarantees to generate feasible schedules for any chromosomes and aims at using genetic algorithm to solve some kind of large JSSP with reasonable solution and reasonable computation time. The approach has been tested with three sets of benchmark JSSP. Experimental results show that the GA with MORS (MORS-GA) can solve the benchmark JSSP of the type mentioned above to optimal or near-optimal with simple GA-operators and fewer objective evaluations. Compared with other GA methods, MORS-GA is shown to be a competitive and promising approach for solving this kind of JSSP.