Gradual Improvement of Benchmarking in Data Envelopment Analysis Using Gradient Line Method

Data envelopment analysis (DEA), firstly, checks whether decision making units (DMUs) are efficient or inefficient and then it introduces a benchmark for inefficient DMUs. This benchmark is of significant importance for managers and decision-makers. There are different methods for benchmarking one of which is the gradient line method. This method has a major problem which is that the benchmark introduced by this method is not always Pareto efficient. Having given an example, this problem is commented on in this article. On the other hand, the application of gradient line is effective on gradual improvement of efficiency because the introduced equation is in such a way that for reducing a certain amount of inputs, the largest expansion is given to outputs. Finally, we demonstrated that by using gradient line in gradual improvement method, there is no need any more to ask the managers for improvement bounds of inputs and outputs in any level and it is enough for the manager to state the highest efficiency improvement amount he expects in each step.