| Semiconductor manufacturing is both a technologically and procedurally complex problem. It is set within a rapidly changing and extremely competitive environment where a premium is placed on reliability and predictibility. What is needed is a new modeling methodology for the development of modeling and simulation tools for this domain.;The methodology developed in this work, MODES, which stands for Modeling Object-Oriented Dynamic Extensible Systems, applies a four-part methodology to building a successful tool. First, an object-oriented framework for representing the problem is developed--i.e. an appropriate collection of objects and methods. Second, the dynamics of the system are properly modeled via simulation. Third, the system architecture is designed to support both graphical and procedural specification and control. Graphical specification allows rapid innovation and interpretation by the user, while procedural elements are used to capture those elements that are difficult or inefficient to capture graphically. Fourth, the system is procedurally extensible in such a way that extensions have complete internal access to the system.;A generic kernel has been developed to support applications being developed using the MODES methodology. The MODES kernel supports a semantically consistent user interface, persistent objects in a virtual object storage system, and object-oriented database, the external procedure interface, and a framework for simulation. The MODES kernel has been used to develop a tool for semiconductor manufacturing modeling, analysis, and simulation, SPDS, the Semiconductor Production Design System. The MODES methodology and kernel, and SPDS are presented. Finally, future extensions and applications, including a tool for semiconductor technology development are presented. |
