Controlling the normalized weights in DEA Models

Data Envelopment Analysis (DEA) is a linear programming technique for measuring and evaluating the efficiency of similar DMUs. As classic DEA models cannot control the weights of inputs and outputs, sometimes, the derived weights from DEA models may not be acceptable by experts. This paper suggests a procedure to incorporate the importance ratios of inputs and outputs, in a DEA model. In fact the experts weights are used in a DEA model as a common set of normalized weights to make the result more acceptable. This common set of weights is added to each DEA model in the form of normalized weight restrictions. An example is used to illustrate this approach.

Keywords:

Data envelopment analysis , a common set of normalized weights , weight restrictions