具有边际收益递增特性的数据包络分析模型

针对传统数据包络分析(DEA)模型不适用于生产可能集前沿面具有边际收益传递特性的实际情况，给出了通过变换函数将不具有凸性的生产可能集变换为满足凸性的生产可能集，再用DEA模型求解，最后用数学反变换求解生产可能集前沿面的方法．变换函数的求取是先构造一系列备选变换函数并分别用DEA模型求解，再判定所取变换函数是否合适，最后筛选出最满意的变换函数．给出了单个投入变量全局及局部边际收益递增时变换函数的求取方法，并通过算例说明了求解过程．结果表明，此方法可用于多输入、单输出的投入产出分析，对变换后的样本集使用DEA模型可得到一个单元的合理评价结果。

A DEA Model Characterized by Increasing Marginal Revenue

Aiming at the fact that the traditional data envelopment analysis(DEA) model is inapplicable to the feasible production set with increasing marginal revenue production frontier, this paper proposed a new method which transforms the feasible production set without convexity to the one with convexity by transform function, then gets solution by DEA model, and gets the result of production frontier of feasible production set by reverse transform function at last. To get the solution to transform function, first prepare a series of transform functions and get solutions by DEA model separately, then determine whether they are appropriate, and screen out the best one. The paper also proposed the method for getting solution to transform function with single input variable when the whole marginal revenue or the local increases. As a result, the method can be used to analyze the devotion and output with multi-input and single output, and can get a reasonable estimate result by transformed sample set using DEA model.