Solving 5x5 Job-Shop Scheduling Problem Using Genetic Algorithm

The job-shop scheduling (JSS) is a production schedule planning for High mix & low volume systems with many variations in requirements. They cannot be formulated as a linear programming and no simple rules or algorithms yield to optimal solutions in a short time. In job-shop scheduling problem (JSSP) environment, there are n jobs to be processed on m machines with a certain objective function to be minimized. JSSP with n jobs to be processed on more than two machines have been classified as a combinatorial problem. JSSP is known to be NP-Hard problem and near optimal solution is possible by heuristics. In this paper, genetic algorithm is used to find the feasible solution set for 5 x 5 JSSP with total flow time taken as the objective function. The processing time & job data for the problem is taken from a glass manufacturing company, which manufactures glass equipment for pharmaceutical company on received order. Population creation, encoding, decoding, selection, recombination (crossover), mutation and reinsertion functions are developed using MATLAB software. The code is run for various generations and solution set containing different sequences with same minimum flow time is presented. The decoding of sequence to schedule is also presented. It is observed that by operation based representation, decoding of sequence to schedule is suitable for JSSP.