A variable weight approach for evidential reasoning

• 作 者: CHANGLei-lei[1]LIMeng-jun[2]JIANGJiang[1]
• 机 构: College of Information System and Management, National University of Defense Technology, Changsha 41
• 出 处: 第八届全国青年岩土力学与工程会议暨青年华人岩土工程论坛
• 发表日期: 2013年期

　　A variable weight approach was proposed to handle the probability deficiency problem in the evidential reasoning (ER) approach.The probability deficiency problem indicated that the inadequate information in the assessment result should be less than that in the input.However, it was proved that under certain circumstances, the ER approach could not solve the probability deficiency problem.The variable weight approach was based on two assumptions: 1) the greater weight should be given to the rule with more adequate information; 2) the greater weight should be given to the rules with less disparate information.Assessment results of two notional case studies show that 1) the probability deficiency problem is solved using the proposed variable weight approach, and 2) the information with less inadequacy and more disparity is provided for the decision makers to help reach a consensus.