| The primary objective of this study is to develop a prescriptive methodology for assessing the total value of any manufacturing expert system project. The study explores in-depth the implementation issues which have been found to affect management perceptions on success or failure for these types of projects.;In order to address the general issues, and to begin to develop the prescriptive methodology of expert system evaluation, it is shown that: 1) Expert system project goals must be well-defined, resources must be committed up front, and management support is a necessary first step toward successful implementation. (2) The acquisition of knowledge must evolve into an orderly, well-structured undertaking; all sources of knowledge and data need to be considered for strategic implication in the project. (3) Fundamental knowledge engineering techniques must be followed in the transformation of the synthesized knowledge base into the expert system prototype, which must then he verified from two perspectives--expert acceptance of each of the individual pieces of knowledge and logical acceptance of the entire knowledge base. (4) The expert system must be validated by proof that it applies to the environment for which it was designed, and that it meets specifications for the target process. (5) Appropriate evaluation techniques must be chosen only after specific parameters have been defined for a comprehensive evaluation of the project.;Through systematic rejection of the null hypotheses in the study, a prescriptive methodology for the evaluation of expert system projects is presented. This prescriptive methodology is derived from a synthesis of evaluative opportunities and critical project parameters as defined throughout the work. A discussion of the findings of this research project leads into a statement of future research needs. |
