Fuzzy Shared Input Two-stage DEA And Its Application In Logistics Enterprise Performance Evaluation

As an important part of the national economy,the logistics industry played a critical role in the development of the society.Besides,severed as one of the leading forces in the logistics industry,improving the efficiency of the logistics enterprises not only beneficial to the long-term and stable development of the enterprises,but also contributes to promote the development of logistics industry.However,if we want to promote the performance of logistics enterprises,we should make a scientific and accurate evaluation to these enterprises.Among the approaches to evaluate the corresponding performance,Data envelopment analysis(DEA),as a nonparametric method for measuring the relative efficiency of homogeneous decision making units,possesses many advantages,such as objectivity,no need to determine the frontier function and so on,and has been widely used in the evaluation of enterprises’ performance.The traditional DEA,however,treats the DMU(logistics system)as a “black box”,ignoring the inner efficiency of production systems.Although the traditional network DEA has the capacity to evaluate the efficiency of the sub-process,but could not evaluate the system with shared resources.In addition,due to data fluctuations or missing and other reasons,the actual data we obtained is often inaccurate or vague.Based upon the above analysis,this paper proposes an approach,which extends the two-stage DEA model with shared resources,according to the possibility approach for fuzzy DEA which proposed by Lertworasirikul et al.(2003),into a fuzzy scenario.Then,we construct an evaluation index system of the logistics enterprises with two-stage production system by analyzing the operation information of the listed logistics enterprises in China.What’s more,the paper also evaluates the logistics enterprises with two different models respectively,that is the two-stage DEA model with shared resources among crisp and fuzzy scenarios.The main conclusions we obtained are as follows:1.during the running period,and taking 23 water transport listed logistics enterprises as a sample,the results show that the overall efficiency of the transport logistics enterprises in China is less than 0.7,with more than 30 percent improving space,and the gap between different enterprises is widening.2.from the perspective of the sub-stage efficiency of the sample logistics enterprises,its profitability efficiency is higher than the efficiency of the marketability stage,the trend of the marketability efficiency is almost synchronous with the overall efficiency,and the principal reason to cause the inefficiency of enterprises is the inefficiency of marketability efficiency.3.there are only six companies among the sample belong to the “high profitability efficiency and high marketability efficiency” pattern,most of the enterprises are deviated from the optimal region,and the improvement room of the overall efficiency is relatively large.4.based on the annual data of the sample logistics enterprises,we also find that the overall efficiency and the corresponding sub-stage efficiency will decline along with the increasing of the possibility level,and these results confirm our several theorems proposed in the paper.5.the ranking discrepancy is not significant between the crisp model and the fuzzy model,and the ranking information from the proposed method is basically fluctuated on the basis of the crisp model.However,compared with the crisp model,the fuzzy two-stage DEA model with shared resources has several advantages,for example,it can give us an evaluation with data loss or uncertainty,besides,it can also,to some extents,reflect the risk preference and utilize the experience of decision makers.