Four principal investigators are leading the project:

• Michael C. Fu
• Jeffrey W. Herrmann
• Steven I. Marcus
• Gary W. Rubloff

The objective of the IPDPM project is to develop new operational methods that enable efficient manufacturing operations throughout the wafer fab life cycle.

Dynamics are ubiquitous in semiconductor manufacturing. The following types of dynamics affect the entire factory:

• Market forces
• Factory initiation and life cycle
• New roadmap nodes
• Technology shrink and shift within nodes
• Multiple product lines and ASICs
• Yield and reliability crashes

At the tool or unit process level, other types of dynamics occur:

• Process re-centering for integration and yield learning
• Tool modifications and upgrades
• Process requalification after preventive maintenance or crash
• Multiple product lines and ASICs

Integrating product dynamics and process models into operational methods helps one answer the following types of questions:

1. How can we rapidly and readily optimize factory operations in the presence of technology shrink, shift, and changes in product mix?
2. How can we rapidly, readily, and concurrently optimize operations and the process in the presence of changes in process recipes and product mix?

The IPDPM project is proceeding along two thrusts: integrating product and market dynamics into operational decision making, and incorporating process and cluster tool models into simulation models of manufacturing operations.

The first thrust includes finite-horizon, Markov decision process models that incorporate transient fab characteristics and that support decisions throughout the wafer fab life cycle. Such models handle the product market's complex dynamics and support the technologies that enable efficient manufacturing for the right products at the right time.

The second thrust integrates manufacturing operations models and manufacturing process models so that decision-makers can predict how process improvements and changes affect the wafer fab's production objectives (thoughput, cycle time, and cost, for instance). The modeling and sensitivity analysis techniques incorporate response surface models, which describe manufacturing processes, and simulation and scheduling techniques, which evaluate the manufacturing system.

Just as concurrent engineering yielded better product designs by teaming design and manufacturing engineers, the IPDPM project aims to improve wafer fab operations by teaming engineers and managers throughout the wafer fab: those who design and control manufacturing processes, those who manage manufacturing operations, those who design and build wafer fabs, and those who design and introduce new products. This project's results will yield specific insights into the structure of optimal operational policies and into the sensitivity of operational decisions to underlying process parameters. Operational and factory integration personnel can use these insights to suggest and justify changes to operational policies and manufacturing processes.

---

Papers

• Shalabh Bhatnagar, Michael C. Fu, Steven I. Marcus, Optimal Multilevel Feedback Policies for ABR Flow Control using Two Timescale SPSA, Technical Report 99-18, Institute for Systems Research, University of Maryland, College Park, 1999.
• Jeffrey W. Herrmann, Niranjan Chandrasekaran, Brian F. Conaghan, Manh-Quan Nguyen, Gary W. Rubloff, Rock Z. Shi, Evaluating the Impact of Process Changes on Cluster Tool Performance, Technical Report 99-30, Institute for Systems Research, University of Maryland, College Park, 1999.
• Praveen V. Mellacheruvu, Michael C. Fu, Jeffrey W. Herrmann, Comparing Gradient Estimation Methods Applied to Stochastic Manufacturing Systems, Technical Report 2000-1, Institute for Systems Research, University of Maryland, College Park, 2000.
• Jeffrey W. Herrmann, Manh-Quan T. Nguyen, Sequencing Wafer Handler Moves to Improve the Performance of Sequential Cluster Tools, Technical Report 2000-3, Institute for Systems Research, University of Maryland, College Park, 2000.
• Manh-Quan T. Nguyen, Jeffrey W. Herrmann, Sequencing Wafer Handler Moves to Improve the Performance of Hybrid Cluster Tools, Technical Report 2000-31, Institute for Systems Research, University of Maryland, College Park, 2000.
• Y. He, M.C. Fu, and S.I. Marcus, A Simulation-Based Policy Iteration Algorithm for Average Cost Unichain Markov Decision Processes, OR Computing Tools for the New Millennium, Manuel Laguna and Jose Luis Gonzalez Velarde, editors, Kluwer Academic Publishers, 161-182, forthcoming, 2000.
• S. Bhatnagar, M.C. Fu, S.I. Marcus, Emmanuel Fernandez-Gaucherand, and Y. He, A Markov Decision Process Model for Capacity Expansion and Allocation, Proceedings of the 38th IEEE Conference on Decision and Control, forthcoming, 1999.
• S. Bhatnagar, M.C. Fu, S.I. Marcus, and Y. He, Markov Decision Processes for Semiconductor Fab-Level Decision Making, Proceedings of the IFAC 14th Triennial World Congress, 145-150, 1999.
• S. Bhatnagar, M.C. Fu, and S.I. Marcus, Two Timescale SPSA Algorithms for Rate-Based ABR Flow Control, System Theory: Modeling, Analysis, and Control, T. Djaferis and I. Schick, editors, Kluwer Academic Publishers, 367-378, 1999.
• Jeffrey W. Herrmann, Niranjan Chandrasekaran, Brian F. Conaghan, Manh-Quan Nguyen, Gary W. Rubloff, Rock Z. Shi, Evaluating the impact of process changes on cluster tool performance, IEEE Transactions on Semiconductor Manufacturing, Volume 13, Number 2, pages 181-192, 2000.
• Gary W. Rubloff, Dynamic Simulation: Guiding Manufacturing from Process Mechanisms to Factory Operations, 47th International Symposium, American Vacuum Society, Boston, October 2-6, 2000.
• Laurent Henn-Lecordier, Manh-Quan Nguyen, Brian F. Conaghan, Praveen Mellachuruvu, Rohit Kumar, Jeffrey W. Herrmann, and Gary W. Rubloff, Integrating Process Models, Equipment Logistics, and Factory Flow for Manufacturing Systems Optimization, 47th International Symposium, American Vacuum Society, Boston, October 2-6, 2000.
• Jeffrey W. Herrmann, Brian F. Conaghan, Laurent Henn-Lecordier, Praveen Mellachuruvu, Manh-Quan Nguyen, Gary W. Rubloff, and Rock Shi, Understanding the impact of equipment and process changes with a heterogeneous semiconductor manufacturing simulation environment, 2000 Winter Simulation Conference, Orlando, December 10-13, 2000.

Last updated: December 18, 2000, by Jeffrey W. Herrmann.