A high speed scheduling method by analyzing job flexibility and taboo search for a large-scale job shop problem with group constraints

A job shop problem with group constraints is a problem where job schedule in each line is decided, and jobs dealt with the same process have to be grouped for every process of a line. The minimum and maximum size restrictions exist in the total size of jobs in the group. An evaluation function of this problem consists of an average manufactured time needed for completion of all jobs, a number of group changes, and a number of line rests. This research consists of two parts: rapid generation of an initial solution by heuristics and improvement of solutions. In the part of generation of an initial solution, the flexibility of each job according to an influential degree of a whole plan is defined. In the part of improvement, rules, which can improve one of three evaluation items, are prepared. By application of a taboo search, which can apply plural rules, total evaluation can be improved. The proposed method has been applied to some real large-scale job shop problems with group constraints, and its effectiveness is confirmed.