An Assessment of Expert System Building Tools for Mining Applications

"Solution of most engineering problems requires, in addition to vast ·numerical calculations, substantial use of practical judgement and expertise. With advances in applied mathematics and computer technology, the numerical aspects of problem solution have reached a considerable level of maturity; However; it is only .recently that approaches like the expert systems methodology have been developed to ·incorporate the Judgemental aspects as well. For rapid development and prototying of expert systems, expert system building tools or ""shells"" have become widely available. For the expert system developer, identification of the key characteristics of the problem domain and selection of the right expert system tool are of crucial importance. In this paper, the important attributes required in an expert system shell for use in the development or a mine ventilation planning expert system are discussed. These attributes are general in that most engineering problems in mining would have these requirements. Also, the results of an exercise undertaken to evaluate commercially available shells for application to mining engineering problems are presented.INRODUCTIONExpert systems (ES) are interactive computer programs which use high quality specialized knowledge in some narrow problem domain to solve complex problems in that domain. These complex problems lack structured and/or standard ways of solution and involve substantial use of subjective and experience based knowledge. One of the first expert systems was MYCIN, developed in 1974, in which subjective and heuristic knowledge of expert physicians was used to diagnose infectious diseases and provide antimicrobial therapy (Buchanan arid Shortliffe, 1984). Since MYCIN, expert systems applications have appeared in other areas such as mineral prospecting (Gaschnig, 1982), configuration of computers (McDermott, 1982) and structural engineering (Bennett and Englemore, 1979)."