对W-B乘子置信域约束构造方法的改进及模型

Wong和Beasley针对数据包络分析(DEA)给出的乘子约束方法(W-B方法)常常会导致相应的模型没有可行解问题。指出了该问题出现的根源在于W-B方法在使用权重重要性概念时存在着逻辑不一致,并提出了一种构造乘子置信域(AR)约束的改进方法及相应的DEA/AR模型。改进方法及模型具有能够克服W-B方法所常常出现的某些乘子置信约束之间相互矛盾问题等三方面优点。经实例应用,验证了所提出的改进方法及相应的DEA/AR模型是切实可行的。

A Modified W-B Approach to Construct AR Restrictions for Data Envelopment Analysis and the Corresponding Model

The approach presented by Wong and Beasley(W-B)to construct assurance regions(AR)for restricting DEA multipliers often results in the problem of no feasible solution The reason lies in that there exists logical inconsistency when the approach uses the concept of importance To solve the problem,this paper presents a modified W-B approach to construct AR restrictions The modified approach has three advantages over the W-B method,e g well overcoming the problem of logical inconsistency Applied to a real efficiency evaluation,the modified approach as well as the corresponding DEA/AR model proves to be feasible and reasonable.