Multi-Viewpoint DEA モデルにおける数値実験

概要

論文の詳細を見る
DEA (Data Envelopment Analysis) is well known as method to measure efficiency of DMUs (Decision Making Units) with multiple inputs and outputs where DEA has an efficiency model and an inefficiency model. The former model finds out the frontier which indicates the best practice line and evaluates the relative efficiency score by the aspect only of the strong points. On the other hand, the latter model finds out the frontier which indicates the worst practice line and evaluates the relative inefficiency score by the aspect only of the weak points. However, the evaluation values between efficiency and inefficiency scores are not consistent. Some DMUs may be in the both state of efficiency and inefficiency. Then, it is not easy for an analyst to understand the difference. To unify the evaluation values between two models, this paper proposes Multi-Viewpoint DEA model. In order to clarify how the proposed model works, this paper presents numerical study for an artificial sample data set. Numerical study shows that the proposed model has a feature which introducing viewpoint parameter allows analyst to evaluate DMUs based on their integrated models as well as the original two models.
2009-07-01