Uncertainty Management for Fuzzy Decision Support Systems

TL;DR

This paper introduces a novel method for managing uncertainty in fuzzy decision support systems by representing proposition certainty with intervals and providing algorithms for inference and evidence combination.

Contribution

It presents new techniques for uncertainty management in fuzzy rule-based systems, including interval-based certainty representation and evidence combination algorithms.

Findings

Developed methods for computing certainty of fuzzy propositions.

Proposed algorithms for evidence combination from multiple rules.

Analyzed the approach's relation to existing methods.

Abstract

A new approach for uncertainty management for fuzzy, rule based decision support systems is proposed: The domain expert's knowledge is expressed by a set of rules that frequently refer to vague and uncertain propositions. The certainty of propositions is represented using intervals [a, b] expressing that the proposition's probability is at least a and at most b. Methods and techniques for computing the overall certainty of fuzzy compound propositions that have been defined by using logical connectives 'and', 'or' and 'not' are introduced. Different inference schemas for applying fuzzy rules by using modus ponens are discussed. Different algorithms for combining evidence that has been received from different rules for the same proposition are provided. The relationship of the approach to other approaches is analyzed and its problems of knowledge acquisition and knowledge representation are discussed in some detail. The basic concepts of a rule-based programming language called PICASSO, for which the approach is a theoretical foundation, are outlined.