Issues in Designing Rule-based Systems for Integrated Evaluation

P. G. Chander, T. Radhakrishnan, and R. Shinghal

The design of rule bases is plagued by various anomalies due to lack of structured development, and lack of a formal reference for the acquired knowledge. Often, the lack of a clear link connecting the functional requirements, the design, and the implementation levels of a rule-based system makes it difficult to analyze the rule base in terms of how well it represents the acquired knowledge. This also makes the existing verification and validation (V&V) tools to stand alone and to be isolated from development. Integration of V&V in a rule-based system design and development life cycle is being recommended by contemporary researchers for quality and reliability improvement. However, V&V processes for rule-based systems have been accepted to be non-trivial: methods that are general enough for comprehensive anomaly detection require impractical amounts of computation, and special methods (reduced computation for V&V) lack in their scope and applicability. In this work, we outline the use of a knowledge acquisition strategy called goal specification and its role as a link that connects the functional, design, and implementation stages of a system. We then identify various issues that affect integrating evaluation into development and how goal specification can facilitate handling these issues.

This page is copyrighted by AAAI. All rights reserved. Your use of this site constitutes acceptance of all of AAAI's terms and conditions and privacy policy.