Data envelopment analysis （DEA） is an effective non-parametric method for measuring the relative efficiencies of decision making units （DMUs） with multiple inputs and outputs. In many real situations, the internal...Data envelopment analysis （DEA） is an effective non-parametric method for measuring the relative efficiencies of decision making units （DMUs） with multiple inputs and outputs. In many real situations, the internal structure of DMUs is a two-stage network process with shared inputs used in both stages and common outputs produced by the both stages. For example, hospitals have a two-stage network structure. Stage 1 consumes resources such as information technology system, plant, equipment and admin personnel to generate outputs such as medical records, laundry and housekeeping. Stage 2 consumes the same set of resources used by stage 1 （named shared inputs） and the outputs generated by stage 1 （named intermediate measures） to provide patient services. Besides, some of outputs, for instance, patient satisfaction degrees, are generated by the two individual stages together （named shared outputs）. Since some of shared inputs and outputs are hard split up and allocated to each individual stage, it needs to develop two-stage DEA methods for evaluating the performance of two-stage network processes in such problems. This paper extends the centralized model to measure the DEA efficiency of the two-stage process with non split-table shared inputs and outputs. A weighted additive approach is used to combine the two individual stages. Moreover, additive efficiency decomposition models are developed to simultaneously evaluate the maximal and the minimal achievable efficiencies for the individual stages. Finally, an example of 17 city branches of China Construction Bank in Anhui Province is employed to illustrate the proposed approach.展开更多
基金Acknowledgments The authors thank the editors and two anonymous referees for their helpful comments and suggestions that substantially improved the quality of this work. This research has been supported by grants from National Natural Science Foundation of China （71224001） and China Postdoctoral Science Foundation funded project （2015M571135）.
文摘Data envelopment analysis （DEA） is an effective non-parametric method for measuring the relative efficiencies of decision making units （DMUs） with multiple inputs and outputs. In many real situations, the internal structure of DMUs is a two-stage network process with shared inputs used in both stages and common outputs produced by the both stages. For example, hospitals have a two-stage network structure. Stage 1 consumes resources such as information technology system, plant, equipment and admin personnel to generate outputs such as medical records, laundry and housekeeping. Stage 2 consumes the same set of resources used by stage 1 （named shared inputs） and the outputs generated by stage 1 （named intermediate measures） to provide patient services. Besides, some of outputs, for instance, patient satisfaction degrees, are generated by the two individual stages together （named shared outputs）. Since some of shared inputs and outputs are hard split up and allocated to each individual stage, it needs to develop two-stage DEA methods for evaluating the performance of two-stage network processes in such problems. This paper extends the centralized model to measure the DEA efficiency of the two-stage process with non split-table shared inputs and outputs. A weighted additive approach is used to combine the two individual stages. Moreover, additive efficiency decomposition models are developed to simultaneously evaluate the maximal and the minimal achievable efficiencies for the individual stages. Finally, an example of 17 city branches of China Construction Bank in Anhui Province is employed to illustrate the proposed approach.
