New Super Efficiency Model And Its Application In Energy Efficiency Evaluation

Data envelopment analysis(DEA)is a non-parametric method to evaluate the relative efficiency of the production frontier.It is an important tool to evaluate the performance of multi-input and multioutput decision-making units.Current DEA models might be infeasible and can not rank DMUs fully under the condition of variable returns to scale and negative data.Based on this,a series of researches on the DEA model which can handle negative data are provided.The main works of this article are as follows:In theory,the defects of the existing super-efficiency models based on the directional distance function are analyzed.Then the conditions are explored,with which the super-efficiency model is feasible and meanwhile its optimal value is bounded.Based on these,several new direction vectors are proposed.These direction vectors can make the corresponding super-efficiency models deal with the data set where zero and negative data exist.Numerical examples show that the proposed models are feasible in the case where negative data exist,can rank DMUs effectively.In application,a new super-efficiency model is applied to evaluate provincial energy efficiency of 30 provinces in our country from 2003 to 2015.Empirical results show that our newly proposed model is superior to the current DEA energy efficiency evaluation models.In the actual production and life,it can deal with negative input-output data sets well,and generate specific performance values for the provinces and cities.By comparing and sorting the energy performance of provinces and cities,we analyze the factors that affect energy performance and provide effective recommendations.This research not only can be useful for energy efficiency evaluation,but also can be applied to other industries containing zero and negative data.More importantly,it extends the application range of the existing super-efficiency model.