Electrical Design: A Problem for Artificial Intelligence Research

Abstract

This report outlines the problem of intelliaent failure recovery in a problem-solver for electrical design. We want our problem solver to learn as much as it cen from its mistakes. Thus we cast the enaineering design process in terms of Problem Solvin& by Debussin& Almost-Risht Plans, a paradiam for automatic problem solvina based on the belief thet creation and removal of "buas• is an unavoid1ble part of the process of solving 1 complex problem. The process of localization end removal of bugs called for by the PSBOARP theory requires 1n 1pproach to engineering enalysis in which every result has a justification which describes the exact set of assumptions it depends upon. We h1ve developed a program based on Analysis by Propa11tion of Constraints which can explain the basis of its deductions. In addition to being useful to a PSBOARP designer, these justifications are used in Dependency-Directed Backtrackin& to limit the combinatorial search in the analysis routines. Althouah the research we will describe is explicitly ebout electricel circuits, we believe that similar principles and methoda ere employed by other kinda of engineers, including computer progremmers.