A Modified Support Logic Programming Approach to Uncertainty Handling in Expert Systems

This thesis proposes a new effective uncertainty handling system, called the Modified Support Logic Programming (MOSLOP) system stemed from the Support Logic Programming (SLOP) system which was proposed by Dr. J. F. Baldwin [1,2]. The original SLOP system consists of an interpreter and a translator. The interpreter runs the SLOP directly while the translator compiles these programs into pure Prolog program. The MOSLOP system is an advanced SLOP system, which uses new forms of support logic pair and modified support logic rules. In the new MOSLOP system, the support pair is made of belief function and the uncertainty measure instead of belief function and plausibility function. Modified Support Logic Programming is the newest way to study the uncertainty measure for expert systems. Several advantages of MOSLOP are explored in this thesis: (a). It is clear and simple to observe and discriminate the performance in an expert system. (b). It simplifies the analysis for a complicated system which is the composition of several smaller systems. (c). It can reduce many operations in the case of missing system parameters. The program for generalizing the MOSLOP system is written in C language using Microsoft C so that it is easy to use