D-S证据理论与信息熵结合的新算法  

An Algorithm Combining D-S Evidence Theory with Information Entropy

作　　者：康健[1] 李一兵[1] 谢红[1] 林云[1] 

机构地区：[1]哈尔滨工程大学信息与通信学院,哈尔滨150001

出　　处：《弹箭与制导学报》2011年第1期197-200,共4页Journal of Projectiles,Rockets,Missiles and Guidance

基　　金：国家自然科学基金(60703090);哈尔滨市科技创新人才研究专项基金(2008RFQXG025)资助

摘　　要：D-S证据合成公式存在着一些不足,针对在计算冲突过大问题上容易出现与人们正常判断标准相悖的问题,文中提出了一种基于信息熵理论的新方法,利用证据数据焦元的差异度来取代证据理论中的冲突因子,结合信息熵理论重新确定焦元所占的比重,对证据进行加权处理,在加权处理过程中应用到指数熵的概念,解决了计算中容易出现不收敛的现象。通过与几种算法进行比较,得出文中算法在处理证据冲突、一票否决问题上的优越性。D-S evidence synthesis formula has some insufficiencies,a new method based on information entropy theory was presented to focus on the conflict between original theory and people＇s normal judgment when the theory was used to solve problems with oversized conflict. In the paper, a method of using the difference of evidence data to substitute conflict factor in evidence theory was proposed with information entropy theory to determine the proportion of focal elements. During the process of weighting the evidence, the concept of exponent entropy was used to solve the problem of convergence. The method proposed in this article emphasized its superiority compared with other algorithms on matters of processing evidence conflict and onevote veto system.

关 键 词：DS理论 信息熵 焦元差异度 

分 类 号：TP212[自动化与计算机技术—检测技术与自动化装置]