As an engineering design evolves, there is a constant interplay of system definition and performace analysis. A successful computer-aided design system must, therefore, not only support tools for convenient interactive three-dimensional geometry definition and modification, but also provide algorithms and computational resources for rapid system simulation. These critical issues of three-dimensional geometry definition and high speed computation are addressed in a unified fashion by the Geometry-Defining Processor to be explored under this research grant. Geometry-Defining Processors, or GPDs, are microprocessors housed in physical geometric packages that can be manually assembled or reconfigured. An individual GPD has information on its own parameters, is able to communicate with neighboring GPDs, and is capable of performing independent numerical computations. An assembly of GPDs is able to define both a three-dimensional system geometry and, operating as an optimally connected parallel processor, efficiently solve the associated continuum partial differential equations needed for design evaluation. Combining simplicity-of-use with powerful computational capabilities, the GPD design system should prove useful in a broad range of engineering applications.
