Research On Non-monotonic Logical Systems In Commonsense Reasoning

Non-monotonic reasoning can be loosely described as the process of drawing conclusions, which may be invalidated by new information. Owing to its close relationship to human common-sense reasoning, non-monotonic inference has become one of the major research topics in the field of artificial intelligence (Al). One of the characters of human mode of thinking is to make decisions by relating two or. more problems to each other. For example, people always associate wet road with raining. In this paper I have attempted to establish a logic system-nonnection? which is to formalize the way of human connecting. This paper provides a detailed study of the connecting logic- connection. Chapter 1 is devoted to a general discussion of the relationship of non-monotonic reasoning to common-sense reasoning. Chapter 2 presents the motivation of establishing the connecting logic. In this chapter the reader will find the syntax and the semantics of the connecting logic. I have verified the completeness theorem and the soundness theorem. Chapter 3 studies the relationship of connecting logic to non-monotonic reasoning and unifies the connecting logic in a modal framework. In chapter 4, Basing on the multiple belief revision, I construct a preferential model which is classical, stoppered, faithful and complete. And the preferential inference operator of the model is a multiple revision operator. Chapter 5 is of implementation problem on R-structure. I establish a variant of tableau system for propositional calculus and then tableau systems for R-structure about a proposition A. Their soundness and completeness are considered. Finally, chapter 6 summarizes the main contributions of the paper and other work of the author during being a doctoral candidate, and discusses some open problems.