According to the reasons of spontaneous combustion in coal seam,we can build up the factors and framework of the spontaneous combustion risk assessment in coal seam,at same time,put up a new way on all-around risk assessment with inexact reasoning.

In order to suit the requirements of blast furnace expert system,an inexact reasoning model has been developed.

The inference engine of the system simulates the real diagnosis of the experts and combines the forward reasoning with inexact reasoning.

Based on the subjective Bayes method, a total probabilistic reasoning model for approximate reasoning is presented.

This paper analyzes the characteristics and limitations of approximate reasoning (based on probability) and nonmonotonic reasoning (based on truth maintenance system).

Study on the uncertain reasoning model of the diesel engine diagnosis expert system;

The certainty factor based model for uncertain reasoning and its calculation method are discussed.

In this paper, we will discuss the certainty factor by using λ additive fuzzy measures, and a kind of algorithm with respect to uncertain reasoning are given.

Uncertain Inference in the CVE-based Intrusion Detection Expert System;

Based on certainty theory, RESTOOL organizes its knowledge base and realizes its uncertain inference engine.

In the system,the method of knowledge representation is genetic rule,inference engine combines uncertain inference and rule inference.

The parameter estimation distribution function was studied by using Bayes probability reasoning method of uncertainty knowledge reasoning in ammunition storage long-term,and changing the experimental method of applying fixed number of storage year to estimate ammunition quality.