| Summary |
The Data Envelopment Analysis(DEA) is a widely used operating performance evaluation method and is often collaborated with Meta-frontier. However, when using Meta-frontier, the users must subjectively divide the data into groups beforehand. For example, the data could be divided according to their locations or industry differences, etc. Sometimes this approach seems to be a bit informal since it has no guidelines. Besides, the data clustering step has completely nothing to do with the operating performance evaluation step. Therefore, this study proposes a Meta-frontier Data Envelopment Analysis Model under Optimal Clustering that allows the users to set the number of groups, and the model would automatically determine the best grouping result and calculate the optimal relative efficiency of each decision-making unit (DMU) simultaneously.
In this study, a genetic algorithm was used to solve the problem and the operational data of Taiwan highway bus operators in 2016 was the research subject. The empirical results show that the model provides a reliable and stable way to assess the relative efficiency among different DMUs. Also, the setting of the group number is very flexible, and the number can be adjusted according to the type or the amount of the data used. Besides, the users can also determine the number of the groups eventually after considering all the assessment results under different group number.
Take four groups as an example, the evaluation results show that: the companies in the first group are good at controlling the number of employees. Zhongxing Bus Company produces less output, while Capital Bus Company is better at generating revenue. In the second group, Fengyuan Bus Transportation Company is the most inefficient. Moreover, all the companies are not good at controlling the number of employees. The companies in the third group have clear differences between their efficiency score. Among them, Chang Hua Bus Company is the most inefficient, while Taoyuan Bus Company and All Day Bus Company are the most efficient. Furthermore, all the companies are not good at the use of vehicles and relatively appropriate for the use of oil. Among the companies in the fourth group, San Chung Bus Company is the most inefficient. Besides, most of them are not good at controlling the number of vehicles. Taipei Bus Company and New Taipei Bus Company are in a state of increasing returns to scale and may consider increasing their operational scale. Overall, every company has differentweaknesses and shortcomings. Some of them have a waste of inputs or a shortage of outputs, while others require adjustments in the operational scale. |
