Ed P. Andert Jr.
The state-of-the-practice in Knowledge-Based System (XBS) verification and validation (V&V) three major problems: developers are unaware of the need for and basic issues of V&V, tools are generally unavailable to aid developers, and research has concentrated primarily on static anomaly detection which alone is inadequate for validation of real-world KBSs. This abstract proposes an approach to improving the situation that concentrates on two key areas. The f'LrSt is exploring a common knowledge representation and KBS execution environment to be augmented with integrated automated V&V tools. The goal is to develop the integrated environment as an open system for wide dissemination to developers. The second improvement is to develop innovative automated dynamic analysis techniques to augment current static anomaly detection capability.