Design of mineral deposit prediction expert system based on GIS and answer set programming

Abstract

The past mineral deposit prediction expert systems cannot support nonmonotonic reasoning. This article introduces answer set programming, which is a well-known formalism in artificial intelligence, to design an expert system for mineral deposit prediction. The system includes GIS spatial database, fact generator, temporary fact library, fact management module, knowledge based management module, knowledge base, reasoning machine, and explanation module. The incomplete deposit knowledge and nonmonotonic reasoning is supported in the system. To demonstrate the feasibility of our approach, we illustrate how to represent deposit information in the form of logic programs and how to make reasoning according to the answer set semantics by an example of the prospecting model of porphyrite iron. The processing of ore deposit predicting is a typical example of nonmonotonic reasoning. As we got more information, we found that the original conclusion was wrong, then we revised our conclusion. Such behavior is supported in our system but is not supported in previous similar systems.