Optimizing The Distributed Production Scheduling Problems Using An Immune-Based Algorithm

Abstract

The challenges faced in the manufacturing industry involve problems that developed in each passing time, particularly in the semiconductor assembly areas. Satisfying customer demands and achieving higher profits, as well as maintaining high productivity have been the most important aspects which attract the expatriate’s attentions in the semiconductor industry. The combined of productivity and flexibility provided by semiconductor assembly areas have elicited variety research efforts for several years. In addition, globalization trends in the manufacturing have encouraged a decentralized effort in the semiconductor assembly industry by implementing distributed production scheduling systems in the production line. With respect to the above mentioned problems, several approaches have been introduced which can be categorized based on the static and dynamic settings of the scheduling. An immune algorithm (IA), which is slightly modified to conform to the manufacturing constraints as well as solving the underlying problem, has been proposed and tested with public data set and industrial data set to establish its effectiveness. The proposed IA algorithm is selected due to its explorative powers of the hyper-mutation operator, solution diversity through its receptor editing operator, and incubation of the memory cell. The results from the experiments was achieved from the proposed IA is effective for the aforementioned problems where the best solution obtained from the public data is between 11% to 19% deviation, while production efficiency for the case studies had obtained within the range of 10% to 66%.