The problem of asking the minimum number of questions in horn clause systems

Abstract

This paper deals with the issue of question asking strategies in expert systems with Horn clause rule bases. A good strategy would ask as few questions as possible to reach a conclusion. This paper describes the development of an efficient and effective heuristic approach which is an extension of the strategy developed by Wang and Vande Vate [1]. Similar to the original strategy, the proposed approach is also organized into two phases. In the first phase, a set of candidate questions is formed as in the original strategy. In the second phase, a question is selected from the previous set. A new question selection rule is used in the proposed strategy. This rule is optimal, given a set of candidate questions. Furthermore, computational results indicate that the new question asking strategy is a highly effective and efficient practical approach.