A DEA-BASED METHOD OF STEPWISE BENCHMARK TARGET SELECTION WITH PREFERENCE, DIRECTION AND SIMILARITY CRITERIA

Abstract

Data envelopment analysis (DEA) is a useful tool based on which inefficient decision-making units (DMUs) can perform a benchmarking process to improve their performance. However, several practical problems need to be addressed in benchmark target selection. One issue discussed in this research. is that it might not be feasible for an inefficient DMU to achieve its target's efficiency in a single step, especially when the DMU is far from the benchmark target DMU. To resolve this problem, various methods of stepwise benchmarking have been proposed. Most of these methods, however, only consider the efficiency score in selecting benchmark targets and ignore various practical aspects that should be considered. In this paper, we propose a new method of stepwise benchmarking based on three criteria: preference, direction and similarity. The first criterion, preference, is used for selecting an ultimate benchmark target; the second criterion, direction is used in selecting intermediate benchmark targets which are located more closely to the improving path; and the third criterion, similarity is used for determining intermediate benchmark targets which are similar to the DMU under evaluation. Considering these three criteria, we develop a method of constructing a more practical and feasible sequence of benchmark targets.