Operations scheduling in job-shops

Abstract

This thesis is in the area of operations scheduling in an automated or semi-automated job-shop system, where parts with various routings contend for service on the resources in the system. The classical job-shop scheduling problem has usually been solved in literature as a machine scheduling problem, neglecting the material handling tasks that effect transfer from one processing machine to another. This has helped in reducing the size of the problem and in decreasing the solution computing time, or in producing a better solution in a given time. In this thesis the complete operations schedule for a job-shop scheduling problem is generated by two approaches in the same solution computing time- (1) by solving the complete scheduling problem directly, and (2) by solving the simplified machine scheduling problem and using its solution to generate the solution to the complete scheduling problem by "adding-on" the material handling activities. The relative performance of the two approaches is studied by testing a suitable hypothesis using results from various experiments. It is concluded that the approach of solving the complete scheduling problem including material handling appears to be better (on average) than solving the truncated problem and generating a solution for the complete problem.