Stochastic DEA

Although DEA offers more advantages than many other statistical approaches, some limitations have to be considered. One important problem is its sensitivity to data. Therefore, a key to the success of the DEA approach is the accurate measure of all factors, including that of inputs and outputs. However, in many situations, such as in a manufacturing system, in a production process, or in a service system, inputs and outputs are so volatile and complex that they are difficult to measure in an accurate way. Thus, some researchers have proposed several models to deal with the data variation in DEA by stochastic models. Sengupta [32] generalized the stochastic DEA model by using the expected value to the stochastic inputs and outputs. Banker [3] incorporated statistical elements into DEA and developed an approach which aims to effect inferences in statistical noise. Many papers (Olesen and Petersen [30], Banker [2], Cooper [12, 14], Land [25]) have introduced chance-constrained programming to DEA in order to accommodate stochastic variations in data. Additional stochastic DEA approaches can be found in Horace [21], Gong [19], Simar [33, 34], and Grosskopf [20].