Mohammad Reza Alirezaee - Department of Mathematics, Iran university of science and technology, Tehran, Iran
Masoud Khalili - Department of Mathematics, Iran university of science and technology, Tehran, Iran

The use of weight restrictions in DEA is a well-established procedure for preventing the DMUs to ignore or overestimating some inputs and outputs in the efficiency assessment. The most popular weight restrictions are assurance regions (ARs), which impose ratios between weights to be within certain ranges. ARs can be categorized into two types: ARs type I (ARs-I) and ARs type II (ARs-II). ARs-I specify ratios either between input or output weights separately, whilst ARs-II specify ratios that link input to output weights. DEA models with ARs-I successfully maximize relative efficiency, but in the presence of ARs-II the DEA models may under-estimate relative efficiency or may become infeasible. In this paper we discuss the problems that can occur when using ARs-II and propose a new nonlinear model that overcomes the limitations discussed. To ensure global optimality the nonlinear model should be solved by the solvers that can find the global optimal solution of nonlinear models like as PATHNLP provided by GAMS.

Data Envelopment Analysis, Weight Restrictions, Assurance regions type II