Optimization of job scheduling in a machine shop using genetic algorithm

As job scheduling involves allocation of jobs to machines to reduce the idle time of machines, the aim of this work emphasises on minimizing the cycle time by using genetic algorithm (GA). Each job has a pre-determined process sequence and the sequences are decided according to metal cutting theory and technological constraints. A modified version of GA known as string GA has been used to get the near optimal cycle time for permutation analysis. An experiment has been carried out with 2 iv 5 resolution to find the significance of five parameters of GA, namely population size, maximum generation, probability of crossing, probability of mutation and crossover operators. Computer runs were carried out with these parameters at various levels and the results indicated that, only probability of mutation, the combined effect of maximum generation and probability of crossing are significant at 10. It is suggested that the minimum values of these parameters be used for scheduling problems.