A robust logic for rule-based reasoning under uncertainty

Abstract

A symbolically quantified logic is presented for reasoning under uncertainty that is based upon the concept of rough sets. This mathematical model provides a simple yet sound basis for a robust reasoning system. A rule of inference analogous to modus ponens is described, and it is shown how it might be used by a reasoning system to determine the most likely outcome under conditions of uncertain knowledge. An analysis of the robustness of the logic in rule-based reasoning is also presented.<>