Abstract
In this paper, we examine the set-up problem at semiconductor assembly and test facilities in a multi-machine, multi-tooling environment. Our primary objectives are to minimise the number of shortages of key devices and to maximise weighted throughput over a 2–5-day planning horizon. When a machine set-up is called for three components must be taken into account: tooling, package size and flow. Each in turn imposes increasing set-up times, ranging from a few minutes to half a day. To balance system efficiency with meeting customer demand, a hierarchical approach is taken. Priority is first given to set-ups that can process (hot) lots that reduce demand shortages. Next, changeover time is factored into the decision. Here, priority is given to the component that takes the least amount of time. Our model determines machine set-ups, lot assignments and lot sequences using a greedy randomised adaptive search procedure. The results indicate that reducing set-up times can reduce hot-lot shortages by up to 6.6% over a two-day period and that following the hierarchy set-up rule while prioritising key device lots can reduce shortages by up to 10%. Moreover, when reentrant flow is taken into account, improvements of up to 42% may be realised.