项目名称: 基于分离的局部信念修正研究及其软件诊断应用
项目编号: No.61262029
项目类型: 地区科学基金项目
立项/批准年度: 2013
项目学科: 自动化技术、计算机技术
项目作者: 吴茂念
作者单位: 贵州大学
项目金额: 47万元
中文摘要: 信念修正是人工智能、数据库理论以及哲学逻辑研究的热点问题。寻找高速有效的算法一直是信念修正研究及其应用的重点。降低信念修正计算复杂性的最有效的方法之一是将计算问题局部化。近年来,基于分离的局部化技术受到研究者们高度重视。通过对命题公式的最细划分, 信念集中非相关原子可被有效的分离出去,从而实现信念修正操作的局部化。该方法亦成功应用于其它相关知识库维护领域。本项目从Horn子句集的最细分离出发,提出一整套基于Horn逻辑的局部信念修正理论,进而将这一理论从命题逻辑推广到一阶Horn逻辑,并探讨一阶逻辑公式集最细分离的存在唯一性,以及计算一阶逻辑及部分特殊公式集最细分离算法。基于这一系统化研究成果,将在Reiter诊断理论中引入基于分离的局部化技术,以实现诊断推理的局部化,从而降低诊断推理过程中几乎无法避免的组合爆炸,以便为解决Reiter一阶诊断理论中的计算复杂性问题提供一种全新的解决途径。
中文关键词: 信念修正;局部化技术;机器人知识库更新;策略推理;协商机制
英文摘要: Belief revision is one of the most important research areas in Artificial Intelligence, Databases and Philosophical Logic. Developing highly effective and efficient algorithms for belief revision operations is a key issue in the research of belief revision and its applications. One of the most efficient approach to reduce computational complexity of belief revision operations is computational localisation. In recent years, the techniques of splitting-based localisation have received considerable attention from the researchers in Artificial Intelligence. Through a process of finest splitting on propositional formulas, irrelevant atoms in a belief set can be effectively separated from the belief set. As a consequence, the computation of belief revision operations can be localised. This approach has been successfully applied to other domains of knowledge base maintenance. This project aims to develop a systematic theory of computational localisation in first-order logic. We will first establish a theory of splitting in Horn belief revision based on the finest splitting technique on Horn clauses. We then extend the theory to the first-order Horn logic. We will investigate the existence and uniqueness of finest splitting on first-order formulas and develop algorithms to calculate the finest splitting of first-order
英文关键词: belief revision;localization methods;update on knowledge of robotics;strategy reasoning;negotiation mechanism