DETERMINING JOB SHOP SCHEDULING USING GENETIC ALGORITHM (GA) TO MINIMIZE MAKESPAN IN AUTOMOTIVE BODIES MANUFACTURING OF PT. XYZ

Abstract

PT. XYZ as one of manufacturing company that produces spare part for the automotive bodies, has deal with huge demand from customer that leads to the unbalance of job and machine. This will affect to the delay in some jobs and delivery of the product because the jobs are incomplete. In addition, there is no integrated system for the production scheduling; the jobs orders are determined based on fist come first serve rule. The current system is using semi-active scheduling approach and requires 637 minutes to do 6 jobs with 5 machines. The genetic algorithm model is proposed using numerical computation software after validation and verification are done. Genetic Algorithm (GA) is done through several steps, including initialization, determining objective function, selection (roulette-wheel), crossover (one-point), mutation (order changing), and breeder GA. The current system and proposed model are using semi-active scheduling approach and being compared to show the differences. After 50 generations are obtained, the optimum solution is shown in generation 6 with makespan 597 minutes. Thus, genetic algorithm model is effectively reducing the makespan of the job shop scheduling problem by 10%.