摘要
针对虚拟研究中心伙伴决策过程中信息的模糊性和不确定性,提出了一种基于支持向量机和模糊层次分析法的综合评价模型。首先应用基于支持向量机的分类模型对虚拟研究中心候选合作伙伴数据信息进行分类,经过初步筛选缩小候选合作伙伴集合,再采用梯形模糊数构造评价模型的判断矩阵,选用一致性充要条件检验判断矩阵的一致性,进而推导出评价模型中各评价准则下指标的权重,最后运用梯形模糊数的重心法,完成候选合作伙伴的综合评价值排序,并以雅砻江水电开发虚拟研究中心伙伴选择为例进行分析。研究表明,基于支持向量机和模糊层次分析法模型能减少虚拟研究中心伙伴选择决策的输入量,提高分类精度和决策效率;此外,由于梯形模糊数在层析分析法中的应用,使综合评价模型能有效地结合定性与定量分析,减少人为主观因素的影响,从而优选出最佳的合作伙伴。
To solve some problems caused by the information fuzziness and uncertainty in selecting cooperative partner of Virtual Research Center （VRC）, a comprehensive decision-making model based on Support Vector Machine and Fuzzy Analytic Hierarchy Process （SVM-FAHP） was proposed. Firstly, the SVM classification model was con- structed to classify the data information of partners and the alternative set was reduced after the primary evaluation. Secondly, the judgment matrix was established in trapezoidal fuzzy number and the priority weights of criteria and sub-criteria were calculated. Then, the consistency of judgment matrix was tested in the necessary and sufficient condition of consistency. In the end, the alternatives were ranked with the integration of the fuzzy gravity method. Finally, the partner selection for the Yalong river hydropower development VRC was presented as an example. It was showed that SVM-FAHP model could decrease the inputs of partners＇ decision-making of VRC and improve classification precision and decision making efficiency. In addition, the application of the trapezoidal fuzzy number enabled the model to combine qualitative analysis with quantitative analysis effectively, reduced subjective influences and selected the best cooperative partner objectively.
出处
《计算机集成制造系统》
EI
CSCD
北大核心
2009年第11期2266-2271,2279,共7页
Computer Integrated Manufacturing Systems
基金
国家自然科学基金雅砻江水电开发联合研究基金资助项目(50679098)
科技部水利部公益性行业科研专项资助项目(200701008)~~
关键词
支持向量机
模糊层次分析法
虚拟研究中心
伙伴优选决策
水电开发
support vector machine
fuzzy analytic hierarchy process
virtual research center
partners optimum decision-making
hydropower development
