Network hierarchical DEA with an application to international shipping industry in Taiwan

Data envelopment analysis (DEA) has been proved to be a powerful approach for measuring the performance of decision making units (DMUs). However, the conventional black-box approach tends to neglect the internal structure of the components and the possibility of having different network structures of DMUs. In reality, DMUs can have complex networks that are in forms of parallel or serial structures and hierarchical processes. For example, in the international shipping industry, the operational tasks can be divided into two stages: supervising the ship dispatch and controlling the work time in the port, which jointly constitute a two-stage operating network structure and each contains its own embedded hierarchical structure. Herein, this study intends to propose a new network hierarchical DEA approach to evaluate the performances of such two-stage structure that embedding the hierarchical structures. Data collected from the Maritime and Port Bureau (MOTC) in Taiwan (2017) is used to validate the reliability and efficiency. The result indicates the effectiveness of the model and provides meaningful implications for the international shipping industry.