Performance Evaluation And Ranking Of Organizational Units

Organizational units refer to as combined groups of the society and material with a certain structure in order to achieve specific goals consciously. In DEA theory, an organizational unit is also called a DMU. As a powerful tool for measuring the performance of DMUs, DEA has developled vigorously in the past30years since it was founded in1978. This method does not need to preset production functions, and also need not to pre-assume the weights for inputs and outputs in prior. Therefore, the evaluation results prevent the interference from some subjective factors. However, with the extension and development of content of DEA, the traditional DEA method also gradually shows some limitations, such as unable to further distinguish efficient DMUs under traditional DEA models, the resulting weights are not unique and the overvalued efficiency score for the evaluated DMU, and so on, which weaken the application ability of DEA in the actual problems.Based on the analysis of the DEA method, the issue of performance evaluation and ranking of DMUs is investigated in the current paper. In the body of this paper, we look into the issue from different situations, including undesirable outputs, common weights, super-efficiency and SBM and carry out a series of theoretical research and discussions on application cases. The main contents of this paper include:(1) It extends the index of "Maximal Balance Index" to the situation where DMU with undesirable outputs is considered from the view of profit competition. Meanwhile, two types of models are built in order to evaluating and ranking DMUs. Through factural examples, it was proved that DMUs can be ranked effectively and uniquely using the maximal balance index.(2) It constructs two models based on ideal DMU and anti-ideal DMU from two aspects of optimism and pessimism, and determines an optimal set of common weights through a secondary goal function model respectively. So, evaluating and ranking DMUs with a set of common weights could be realized.(3) It builds a non-radial super-efficiency model by introducing the input preferences of the decision makers from the view of practical significance. Compared to the previous radial super-efficiency model, the proposed model has flexible option of selecting the amount for each input. It also studies and analyzes the on-court efficiency and ranking change of26basketball players in2014-2015NBA regular season using the non-radial super-efficiency model incorporating input preferences.(4) It proposes an integrated model with the factors of input saving and output surplus on the basis of slacks-based measure. The integrated model can guarantee that the projection on the frontier for the evaluated DMU is pareto-efficient and reveal the causes of the resulting efficiency. Next, this chapter measures and analyses E-government performance of31provincial government websites in China with slacks-based efficiency measurement.In the end, the main conclusions of this paper are summarized. And some disadvantages in this paper and suggestions for future research are put forward.