基于DEA模糊窗的测试软件评价研究

Testing software evaluation based on DEA fuzzy window

在测试软件的评价问题中,由于指标值、指标权重关系都具有主观模糊性等问题,将模糊数据包络分析(data envelopment analysis,简称DEA)引入评价过程.为改进现有模型在权重关系信息的模糊性处理方面存在的不足,提出了基于模糊DEA的测试软件改进评价模型,并提供了该改进模型的求解算法.所述方法首先将主观指标值模糊化,转化为模糊数;然后以模糊权重信息为约束条件,构造含有偏好信息的模糊窗,建立只有输出指标的模糊DEA模型;最后,改进模糊DEA模型经由模糊数截集的方法,转化为确定性DEA模型并进行求解.通过实例分析验证了该方法的一致性和可行性.

In testing software evaluation,since index value and index weight relationship were subjective fuzziness,the DEA was introduced into the evaluation process,so as to improve the fuzzy processing deficiencies of the weight relationship information,then an improved test software evaluation model was advanced based on fuzzy DEA with the algorithm given.In this method,firstly the subjective indexes were transformed into fuzzy numbers,then constraints were also turned to fuzzy weight information numbers as a preference information fuzzy window,furthermore a fuzzy DEA model with fuzzy window was established with only the output indexes.Finally,the fuzzy DEA model was transformed into a normal definite DEA model and solved by using cut sets of the fuzzy numbers.An application of this method was given for testifying its consistency and feasibility.