Abstract
A multi-head test machine consists of one central process unit(CPU) and multiple test heads. At a particular time there may be several lots of different products connected to the test heads of a test machine. The product combination will affect the processing time of lots. This study considers the factors of product combinations, sequence- dependent setup times, and the differences between completion times and due dates. Then, define the configuration index value and lot index value, and use these two index values to modify the search processes of tabu search, simulated annealing, and genetic algorithm. Using the various search methods solves the problem with the objective of minimum average tardiness. This study uses the Apparent Tardiness Cost with Setups (ATCS) priority rule to calculate an initial solution. By computation experiments, we find that the tabu search using the lot index value is an efficient algorithm to solve the problem of multi-head test machine scheduling.