Explanation-Based Reasoning in Decision Support Systems

Abstract : In Navy command and control systems there are increasing problems with situations in which human technicians must make quick decisions based on uncertain information. The SABER tool is being developed using an explanation- based reasoning approach to assist in making such decisions. Explanation-based reasoning is modeled after an explanation-based decision making process through which people are believed to perform some decision making tasks. SABER constructs and evaluates alternative explanations to account for input data that may be incomplete or inconsistent. The explanations are evaluated according to three criteria: simplicity, completeness, and significance. Users are given all of the explanations in ranked order. SABER is designed to allow users to easily change the data and to modify the rankings of the explanations. This emphasis on modifiability effectively enables users to train SABER. Cognitive modeling, Displays, Stress, Decision making, Man-machine interface, Team decision Making, Decision support, Performance