Priority mix planning for semiconductor fabrication by fuzzy AHP ranking

Abstract

With today’s keen competition, semiconductor market has changed from producer-oriented to customer-oriented. To be successful, companies need to consider both customer satisfaction in demand and the ultimate profit goal of companies. Semiconductor fabricators today must face an environment with multi-product types and multi-priority orders. Since semiconductor fabrication has a very complicated production process, the production planning of different products types and priority levels is an even more difficult task to experts. The objective of this study is to construct an analytical approach under a fuzzy subjective judgment environment, in which fuzzy analytic hierarchy process (AHP) method with entropy weight is utilized to deal with uncertainty, to generate performance ranking of different priority mixes. The results provide guidance to experts in a fab regarding strategies for accepting orders with the consideration of manufacturing efficiency in the aspects of product, equipment efficiency and finance.

Introduction

In the construction of a wafer fab, a very high capital investment in plant and equipment, from $US 500 million to 1 billion each, is required. In addition, wafer fabrication involves the most complex manufacturing system among all the manufacturing industries. Production planning of a semiconductor fabricator is very difficult due to its distinctive complexities in the manufacturing process. The process may consist of 300–500 sequential processing operations and a flow time of usually more than twenty days. The production planning and scheduling for the complex manufacturing processes are a challenge due to the factors such as complex product flows, random yields, diverse equipment characteristics, equipment downtime, production and development in shared facilities, data availability and maintenance (Atherton and Atherton, 1995, Uzsoy et al., 1992). On top of that, different operations may require the use of the same process equipment, and this is the so-called re-entry characteristic. Thus, a decision made to assign an operation to run on a machine will affect the future demand on this machine, and affect the smoothness of the production flow.

Product mix determination is one of the core problems in current semiconductor production planning system. Different products require different manufacturing processes, and the requirements of setup may also be different. The process plans of products can range from very identical to being extremely distinctive depending on the types of products. The greater is the difference among the process plans, the more diverse are the loading demand and batch difficulty on the factory. In order to best utilize a current fab, a proper selection of product mix is necessary.

Multiple priority levels of orders are usually apparent in wafer fabrication, and higher priority must be given to some urgent lots in order to be competitive and to satisfy customers’ demand of accelerating the speed of products entering into the market. When a full loading policy is not required for batch machines and machines are thus not fully utilized, processing a higher priority order can result in machine capacity loss and a elongation of production cycle time of normal orders is resulted. As the lots with higher priority increase, the variation in shop floor performance will increase and the system throughput will reduce.

In conclusion, product and priority mix has a tremendous impact to the production system, and product mix with different multiple priority lots has different and great influence on the system and pose a great challenge to wafer fabrication. Many performance factors such as cycle time, WIP level, throughput, bottleneck utilization rate will be affected. Organizing the available data is a complicated task, and different people involved in the decision-making may have different opinions on these performance factors. In addition, uncertain thinking process of human beings is present. Therefore, this research proposed a fuzzy AHP model with entropy weight concept to deal with multiple performance factors and to evaluate which product and priority mix can provide a more stabilized production environment and a better overall outcome for a wafer fab. The proposed model can be followed by administrators to determine the most suitable priority mix and can provide a guidance regarding strategies for aggregate planning so as to improve manufacturing efficiency for a fab.

This paper is organized as follows. Section 2 goes over the key concepts of priority levels in production, entropy method and fuzzy AHP. Section 3 presents the methodology and algorithm. Section 4 applies fuzzy AHP based on entropy weights to the evaluation of the efficiency under different priority mixes. Some conclusion remarks are made in the last section.