One of the more successful applications of artificial intelligence techniques has been the development of programs that have come to be known as rule-based expert systems. These programs encode knowledge about specialized problem areas in the form of sets of if-then rules, typically obtained by interviewing people who are specialists or experts in those problems. The rules can then be used by the program to solve similar problems. The modular structure that results allows incremental development, leading to performance that continually improves as the rule base is expanded.

Most of the rule-based expert systems that have been developed to date have been designed for problems in medical diagnoses. However, among several efforts that are relevant to the petroleum industry, a program called Prospector has been developed for the U.S. Geological Survey for certain problems in hard-rock mineral exploration, and a program called the Dipmeter Adviser has been developed by Schlumberger for the geological interpretation of dipmeter data. This presentation will describe the basic principles behind all such systems and will summarize the current state of the art.

End_of_Article - Last_Page 1704------------