PEMES — A Petroleum Exploration Oriented Model Expert System

Abstract

PEMES is a Model Expert System which was designed to suit the particular characteristics of problems in the petroleum exploration domain. This paper describes an approach to using ‘Tasks’ to organize the reasoning process of the system and to integrate numerical calculation with logical inference. It also presents a brief discussion about the logical data structure which is used to deal with the complex objects used for problem solving. The execution of the PEMES system is a procedure of establishing tasks and performing tasks step by step. The work of establishing tasks and performing tasks is done by the System Control module and the Knowledge Interpretation module. The knowledge representation in PEMES is divided into two parts: 1). Task transformation rules, including task decomposition rules, task derivation rules and task ordering rules; 2). Inference methods each of which contains production rules and production-refinement rules. Correspondingly the Knowledge Interpretation module is also divided into two parts: 1). The Task Knowledge Interpretation module. 2). The Domain Inference Method Interpretation module.