In times of increasing global competition and decreasing product and service lifecycles, new business paradigms are emerging as a result of growing pressure to optimize the utilization of critical assets and yet to continuously adapt to fast-changing business environments. Specifically in the semiconductor industry, the landscape for manufacturing operations has changed dramatically: complexity is increasing rapidly – greater product complexity (more masks, more materials, more process steps), and more products with faster ramps – at the same time capital cost of factories is stretching toward $4 billion. The search for high asset utilization has also created opportunities for foundries and contract manufacturers. As features become smaller and wafers become larger, fine-tuning of fab capacity becomes more difficult, and the impacts of bottleneck process failure or sudden changes in product mix are magnified.
In this setting, the 3rd International Conference on Modeling and Analysis of Semiconductor Manufacturing (MASM 2005) was held on 6–7 October 2005 as a forum for the exchange of ideas and best practices between researchers and practitioners from around the world involved in modelling and analysis of semiconductor manufacturing. For the first time, the conference took place outside the United States: close to 200 participants from academia and industry around the world came together in Singapore, Asia's ‘Silicon Island’ with 12 wafer fabs and more than 20 assembly and test facilities in operation, to discuss not only theoretical developments but also applied research and best practice industrial case studies. This special issue of IJPR presents a selection of 14 papers that were presented during the conference.
Improvements at all levels are expected to contribute to future cost reductions and revenue growth in the industry: at the lowest level, improvements to equipment and operator productivity are as important as ever. T.E. Lee et al . addressed the specific issue of wet station scheduling with multiple job flows and wafer handling robots. To overcome the limitations of evaluating Operational Equipment Efficiency for a single tool, a set of Overall Tool Group Efficiency (OGE) has been proposed by C.F. Chien et al .
At the factory level, capital cost reductions and operational improvements are expected to make very significant strides. The complexity of today's operations requires high-fidelity models for decision-making. In this context, P. Wirojanagud et al . have addressed the importance of taking into consideration for workforce management the inherent difference of individual workers. A novel methodology for mixed production goaling and control has been developed by A. Kadir. It resulted in significant performance benefits after successful implementation in several semiconductor manufacturing facilities at Intel. Specifically with regard to 300 mm wafer fab facilities, D. Nazzal and L.F. McGinnis have developed an analytical approach for the estimation of the performance of an Automated Material Handling System which helps to overcome some limitations of other simulation-based approaches in terms of cost and time for model development and simulation execution. The critical issue of delivery performance has also been addressed by A. Gupta and A.I. Sivakumar as well as (specifically for foundry fabs) by K.H. Kang and Y.H. Lee.
At the highest level, supply chain efficiency promises to magnify equipment and factory level advances upstream and downstream in the business. Supply chain configuration in general has been addressed by D. Chiang et al . through a polynomial goal programming model that can accommodate different objectives. The question of how the ‘within-echelon’ challenge of coordinating multiple wafer fabs can be addressed has been investigated by S.H. Chang et al . and B.P. Gan et al . The latter paper also demonstrates successful application of distributed simulation for such complex systems.
To achieve these improvements, in addition to operations research methods, new statistical methods will be required. P. Das and S. Datta presented a study for the evaluation of process performance by using statistical and neural network models. Statistical process control has been widely used in manufacturing industry, and advanced control charts are being studied by many researchers today. In two papers, Z. Wu and Q. Wang as well as J.Y. Liu et al . applied advanced statistical control charts such as CUSUM and EWMA for process monitoring and decision making regarding the status of the process. In addition, J.F. Hill et al . presented a study for dynamic detection of outliers in electronics production.
As evidenced by the quality and diversity of the papers in this special issue, the MASM 2005 conference was truly an international forum for exchange of ideas between researchers and practitioners. The next MASM conference is expected to take place in Germany in the Autumn of 2007. At that time, we anticipate that we will see exciting advances in research and practice for today's semiconductor manufacturing challenges and will be introduced to new and exciting challenges for the next generation of semiconductor manufacturing operations.