Production planning with joint resources usage semiconductor test manufacturing

Abstract

A tester and qualified kits are used simultaneously to test a semiconductor product. One kit consists of six or seven components. Production planning in semiconductor testing is complex due to dependant resources requirement and intricate {product, tester, kit, component} qualification relationships. We formulate a multi-period production planning problem employing mixed integer programming (MIP) technique. We consider multiple matching ratios of jointly used testers and kits, allow kit reconfiguration and optimize capacity allocation at the level of kit components. We also conduct weight coefficient analysis to enable multiobjective optimization. Industrial implementation saves millions of dollars in kit purchasing and many man-hours.