An Inverse DEA Model for Input/Output Estimation with Integer Restriction
K G Shinto1, C M Sushama2
1. Department of Mathematics, Christ College Irinjalakuda, India.
2. Department of Mathematics, National Institute of Technology Calicut, India.
Abstract: Basic DEA model considers real valued inputs and outputs and measures the relative efficiency of similar decision making units (DMUs). If a particular DMU is not efficient, a set of target values for the input/output can be suggested by projecting the DMU to the efficient frontier which will make the DMU efficient. Conventional inverse DEA model is used to estimate the input (or output) for a decision making unit (DMU) when some or all outputs (or inputs) of this DMU is changed. But in a DEA model with integer restriction in input and/or output values, these models are not acceptable if the estimated values are not integers. Rounding off these values may not provide the result in all cases. This paper presents a situation when the input values of an efficient DMU are increased. With these new input values and with the existing output values this DMU may become inefficient. The projection of this DMU into the efficient frontier may not preserve the integer restriction on the output values. The paper presents an iterative MILP model to handle the situation.
Keywords: Data Envelopment Analysis (DEA), Inverse DEA, Mixed integer linear programming (MILP), CRS DEA, Output oriented DEA.
Pages: 187 – 194 | Full PDF Paper