权重信息不完全的仿真可信度指标聚合方法  

Aggregation method for simulation credibility index with incomplete weight information

作　　者：杨明[1] 张忠[1] 方可[1] 

机构地区：[1]哈尔滨工业大学控制与仿真中心,黑龙江哈尔滨150080

出　　处：《系统工程与电子技术》2013年第4期885-888,共4页Systems Engineering and Electronics

基　　金：国家自然科学基金创新研究群体科学基金(61021002)资助课题

摘　　要：针对专家偏好信息为残缺区间判断矩阵的仿真可信度评估问题,提出了一种新的指标聚合方法。首先,给出了区间数直径和偏离度的定义。然后,根据专家所给的残缺区间判断矩阵计算出专家的客观权重,再利用凸组合将专家主观权重与客观权重集结为专家综合权重。接着,结合专家综合权重建立了一个确定指标权重的非线性优化模型,进而提出了一种权重信息不完全时的指标聚合方法。最后,通过实例验证了方法的合理性与有效性。For the problem of simulation credibility evaluation, in which expert＇s preference information is in the form of an incomplete interval judgment matrix, a new method of index aggregation is investigated. First, the diameter and deviation degree of interval numbers are defined. Then, the expert＇s objective weight which is calculated with the incomplete interval judgment matrix and the subjective weight are aggregated into an expert＇s syn- thesis weight based on convex combination. Next, a non-linear optimization model is established with the expert＇s synthesis weight to determine the index weight, and a method for index aggregation with incomplete weight information is further proposed. Finally an example is given to validate the reasonableness and effectiveness of the proposed method.

关 键 词：可信度评估 指标聚合 残缺区间判断矩阵 优化模型 

分 类 号：TH166[机械工程—机械制造及自动化]