Expert Systems Show Promise for Customer Inquiries

This article describes results of an agreement between the North Penn Water Authority in Lansdale, Pa., and the US Environmental Protection Agency, Drinking Water Research Division, Cincinnati, Ohio, to study use of expert systems technology in a water utility. The three-year study has resulted in the development of computerized expert systems whose basic principles should be broadly applicable to other water utilities. The system described in this article, called a customer queiy expert system, is designed to assist administrative personnel in handling customers' inquiries and complaints about water quality. The system has been developed as a prototype and will be placed in field operation for further testing and refinement in the near future. Significant effort, however, is expected to be necessary to move from prototype "proof-of-concept" demonstrations to fully featured systems that are of practical use in the field. Interest has been growing in the use of expert systems for decision making in various fields of science and technology. Expert systems have grown out of research on artificial intelligence (AI), which is concerned with making computers perceive, reason, and understand. Although getting computers to reason and solve generalized problems has met with little real-world success, much progress has been made in getting them to deal with specific, focused problems. The recognition during AI research that knowledge is as important as reasoning resulted in the development of knowledge-based expert systems - programs, that face the difficult challenge of representing and exploiting knowledge to solve difficult problems in specialized areas. One approach that has proved popular is the use of "IF-THEN rules" (also called situation-action or production rules). These rules say that if a certain kind of situation arises, a certain kind of action can be taken.