Preventive maintenance scheduling optimization in semiconductor assembly industry using harmony search algorithm

Highly competitive markets and rapidly changing customer demands have increased the importance of proper resource utilization in the semiconductor industry. Utilizing resources (i.e. machines) is vital in order to maintain continuous flow of manufacturing as well as their efficiency and productivity. However, machine maintenance is being overlooked in most cases of the semiconductor assembly processes. This will greatly affect the effectiveness and productivity of semiconductor industry when machine breakdowns happen. Various maintenance models had been proposed by previous researchers which focuses on a different mode and environment of maintenance in manufacturing. The objective of this study is to maximize the maintenance throughput while satisfying resource constraints. A new model using harmony search (HS) algorithm, slightly modified to suit the manufacturing constraints, was proposed and tested with an industrial datasets to examine its effectiveness. A comparative analysis of both with and without the HS algorithm application is conducted. From the results, the proposed HS algorithm manages to achieve a positive improvement of 10% from an original industrial maintenance schedule, where the maximum maintenance throughput increases from 71.43% to 78.57%.