Semiconductor fabrication facility design using a hybrid search methodology

The use of integrated circuits (ICs) is ever increasing in the past decade and will continue to increase in the foreseeable future. To compete success fully in global markets, semiconductor manufacturing facilities require goo d system performance in terms of high throughput and high yield on wafers w ith smaller line widths and larger diameters, while still maintaining low s ystem costs. The installation of an automated material handling system (AMH S) within a typical bay layout configuration is a viable approach for impro ving system performance as well as reducing operating and capital investmen t costs. Integrated approaches for fab layout and AMHS design generally involve mixe d integer programming (MIP) formulations that are difficult to solve for la rge size problems. This paper proposes a hybrid search methodology based on tabu search and simulated annealing to solve such layout and AMHS design i ntegration problems in a wafer fab. The procedure was computationally teste d for efficiency and effectiveness against an existing MIP formulation. The results indicate that the proposed procedure performs very well in terms o f both solution quality and computation time. It is shown that this procedu re is extendable to other layout and material handling system design proble ms. (C) 1999 Elsevier Science Ltd. All rights reserved.