Ranking Prediction Using Integrated ANFIS and DEA Model for Desired DMU

Data Envelopment Analysis (DEA) is a mathematic technique for evaluating the performance of decision-making units. Regarding the fact that evaluating decision-making units using DEA requires solving the linear programming problem, a method is proposed in this paper, using which there will be no need to use Data Envelopment Analysis Method again to solve the problem of evaluating the efficiency and ranking when new DMUs are added to the previous ones or the inputs or the outputs of any of the existing DMUs are changed. In the proposed method, the ranking and the efficiency of the new DMUs can be defined based on the ranking and the efficiency state of the existing DMUs and using artificial neural networks. In this study, Back Propagation Optimization method is used to train the fuzzy neural system. After the Integrated ANFIS learns reference patterns, it calculates the ranking of the new DMU accordingly, and then, based on the ranking of the new DMU and comparing them to previous DMUs, efficiency or inefficiency of the new DMU can be defined.

کلیدواژه ها:

Data Envelopment Analysis (DEA) ، Artificial Neural Network ، Adaptive Neuro Fuzzy Inference System (ANFIS)