Development of an expert system-based experimental frame for modeling of manufacturing systems

Scope and method of study. The overall goal of this research was to develop a methodology that utilizes expert systems in the experimental frame module of an advanced modeling environment for manufacturing systems in order to provide timely assistance to the problem definition and solver selection decision making processes. Such an experimental frame expert system includes (1) a problem definition knowledge base which is capable of deriving a structured problem from a set of symptoms, and (2) a tool selection knowledge base which is capable of recommending the most suitable tools (analysis and/or optimization) to address the defined structured problem. In the construction of these two knowledge bases, knowledge acquired from multiple experts (from both industry and academia) is utilized.;Findings and conclusions. The primary motivation for this research was the author's desire to contribute to the advancement of modeling of complex manufacturing systems by addressing some of the issues related to the shortcomings of traditional modeling methodologies including the lack of reusability and lack of accessibility. At the highest level, the outcome of this research effort is a new framework within which expert systems are combined with object oriented modeling paradigm to maximize the accessibility of models by non-modeling specialists. Lists of symptoms, problems, and what-if questions which can be commonly found in discrete part manufacturing systems are collected, organized, and presented. Mappings of the collected symptoms to structured problems and the structured problems to a set of tools are also collected, organized, and presented. A proof of concept implementation of this framework in which the experimental frame expert system (EFES) works with the advanced modeling environment (AME) under development in the Center for CIM at OSU was developed under the Smalltalk-80 based VisualWorks 2.0 development environment. This implementation demonstrates promising potential for further research and commercial application of this methodology.