目标一致型两阶段网络SBM效率评价方法

Efficiency evaluation of goal-congruent systems using a two-stage network SBM method

网络SBM模型具有考虑系统内部结构、不要求投入或产出等比例改进等优点,但传统网络SBM模型在测算效率时未考虑内部中间产品或中间产品的改进方向不明确,导致模型存在无法测度子阶段效率、子阶段目标与整体目标矛盾等情况.本文提出一种目标一致型两阶段网络SBM模型,优化了中间产品的改进方向,重新考虑了整体目标函数与子阶段目标函数,使得在提高决策单元投入产出水平时子阶段改进目标与整体改进目标一致,改进过程更加符合实际生产情况,并且该模型整体效率等于子阶段效率乘积,避免了传统加权求和方式带来的主观性和偏向性.将其应用于我国高技术产业创新效率测度中,实证结果表明:目标一致型网络SBM模型的评价结果相较于传统模型更加合理,所给前沿投影能够让无效决策单元达到DEA有效状态.

The network SBM model has the advantages of considering the internal structure of the system and not requiring proportional improvement of input or output.However,traditional network SBM models do not consider the internal intermediate products,or the improvement direction of their intermediate products is unclear when calculating efficiency,resulting in the model being unable to measure sub-stages’efficiency,and the contradiction between sub-stage improvement goals and overall goals.The article proposes a two-stage network SBM model with consistent objectives,which optimizes the improvement direction of intermediate products,redefines the overall objective function and sub-stage objective functions,so that the sub-stage improvement goals are consistent with the overall improvement goals when improving the input and output level of decision making units,and the improvement process is more in line with the actual production situation.And the overall efficiency obtained by this model is equal to the product of sub-stage efficiency,avoiding the subjectivity and bias brought about by traditional weighted summation methods.Apply it to the measurement of innovation efficiency in China’s high-tech industries.The empirical results indicate that the results of the SBM model with consistent objective are more reasonable compared to traditional models,and it provides a frontier projection to enable invalid decision making units to reach the DEA effective state.