基于Rough逻辑粒的不确定上下文推理  

An Uncertainty Reasoning Approach Based on Rough Logic Granular for Context Awareness

作　　者：吕方苏[1] 饶文碧[1] 徐铁城[1] 

机构地区：[1]武汉理工大学计算机科学与技术学院,湖北武汉430070

出　　处：《武汉理工大学学报（信息与管理工程版）》2011年第4期540-543,共4页Journal of Wuhan University of Technology：Information & Management Engineering

基　　金：湖北省重大科技专项基金资助项目(2007DA111);武汉理工大学创新基金资助项目(2010-ZY-JS-013)

摘　　要：针对信息模糊或缺失的不确定上下文推理的难点问题,以Rough逻辑为基础,结合粒计算的思想,利用上下文信息本身的层次性,将上下文划分为底层上下文、高层上下文和服务上下文,据此对Rough逻辑公式进行4个层次的粒划分,计算原子粒的取值,由原子粒的值合成整个Rough逻辑公式的值,最终将每个公式的取值映射到[0,1]区间,由此提出了一种针对模糊不完备上下文的具有良好可扩展性和可维护性的推理方法,并通过实例验证了该方法的可行性。Context reasoning is one of the core technologies in context awareness,and the uncertainty reasoning for context awareness has been a difficulty under condition of information intangibility or information deficiency.Based on Rough logic and granular computing,a new uncertainty reasoning approach for context awareness was proposed.Context information was divided into three parts：low level context,higher level context and service level context.The Rough logic formula was granularly divided into four levels to calculate the value of atomic granule.Then the value of atomic granule was integrated into value of whole Rough logic formula.Each formula value was mapped in interval.A reasoning method with good scalability and maintainability was then put forward for information intangible or information deficient context.The feasibility of the approach was verified by examples.

关 键 词：ROUGH逻辑 不确定推理 上下文感知 粒计算 

分 类 号：TP391.1[自动化与计算机技术—计算机应用技术]