Undesirable inputs and outputs in data envelopment analysis models

Efficiency measurement in Data Envelopment Analysis (DEA) is usually based on the assumption that inputs have to be minimized and outputs have to be maximized. In some situations, undesirable (bad) inputs and outputs may be presented in the production process. In order to improve the performance of an inefficient Decision Making Unit (DMU) the undesirable outputs and desirable (good) inputs should be decreased while the desirable outputs and undesirable inputs should be increased. In this paper, we present a method to deal with such inputs and outputs in non-radial DEA models.