摘要
针对属性权重未知且评价数据为多维时序的信任度排序问题,提出一种基于线性规划理论的信任度排序方法。首先使用线性规划模型确定一维时序下属性权重未知的多个节点信任度排序上下限向量,然后建立多维时序下信任度的最优协调排序模型,再将模型转化为典型指派问题并利用匈牙利算法进行求解,从而得到各个节点的信任度排序。实例分析表明,当最优协调排序模型中距离参数q取低值时,对极值数据不敏感,可以防止少数评价数据突变造成的误评;当q选取高值时,对极值数据较敏感,可以识别受评对象中信任度摇摆不定的潜在不诚实对象。
Concerning the trust degree rank question of which attribute weights are unknown and data are multidimensional, a trust degree rank method based on linear programming theory was proposed. Firstly, the proposed method used linear programming model to make sure the trust degree rank ceiling-floor vector of nodes under the condition of one- dimensional data and unknown attribute weights, and then built the optimal coordinating rank model under multidimensional timing, and then converted the model to the classical assignment problem, and used the Hungary algorithm to solve the model, at last the trust degree rank of each node was obtained. The example analysis shows that, when the optimal coordinating rank model distance parameter q takes low value, the extreme data are not sensitive to the model, the proposed model can prevent evaluation error caused by a few mutation data; when q takes high value, the extreme data are sensitive to the model, the proposed model can identify the potential dishonest evaluation objects of a swing trust degree.
出处
《计算机应用》
CSCD
北大核心
2013年第3期720-722,726,共4页
journal of Computer Applications
基金
国家自然科学基金资助项目(71171198)
中国博士后科学基金资助项目(2012M512132)
湖北省自然科学基金资助项目(2011CDB052)
关键词
信任度排序
最优协调排序
线性规划
多属性决策
匈牙利算法
trust degree rank
optimal coordinating rank
linear programming
muhiple attribute decision-making
Hungary algorithm
