A fuzzy-knowledge resource-allocation model of the semiconductor final test industry

The operations of the semiconductor final test industry are complicated and characterized by multiple-resource constraints that require simultaneous considerations. One of the most challenging production-planning decisions in the industry concerns an efficient allocation of resources that results in high manufacturing performance. Firms in the industry are thus eager to discover resource-allocation knowledge from large manufacturing databases. This study develops a novel model via the extraction of fuzzy-business rules from databases for obtaining resource-allocation knowledge as well as allocating resources efficiently. The proposed model uses both a genetic algorithm to find the best priority sequence of customer orders for resource allocation and, in accordance with the priority sequence of orders, a fuzzy-inference model to allocate the resources and to determine the order-completion times. Experiments showed that the proposed model can significantly reduce task tardiness.