Solving complete job shop scheduling problem using genetic algorithm

Abstract

Scheduling is the key coordinating activity in manufacturing industry. Conventional Job shop scheduling problem (JSSP) draws much more attention than the JSSP with assembly operations. We introduced a concept termed CJSSP (complete JSSP) to extendedly define and explicitly describe it as a basic problem. Our objectives include exploring CJSSP and developing an algorithm to solve it. Since no CJSSP benchmark existed thus far, we adapted one from the benchmark FT10. We worked out a genetic algorithm (GA) with a novel encoding process for it. Computation results illustrate that our algorithm is feasible and effective. Moreover, a near-optimal makespan of 2046 was obtained.