Global job shop scheduling with a genetic algorithm

This paper describes a global job shop scheduling procedure that uses a genetic algorithm to find a good schedule. Unlike previously considered algorithms, this procedure has been implemented in the scheduling system for a manufacturing facility and has led to improved scheduling. This facility is a semiconductor test area. The test area is a job shop and has sequence-dependent setup times at some operations. The concern of management is to meet their customer due dates and to increase throughput. This requires the coordination of many resources, a task beyond the ability of simple dispatching rules. The authors discuss a centralized procedure that can find a good schedule through the use of a detailed scheduling model and a genetic algorithm that searches over combinations of dispatching rules. They discuss the present effort in developing a system that models the shop, creates schedules for the test area personnel, and makes a number of contributions to test area management.