Centralized resource allocation and target setting of a two-stage production process using data envelopment analysis

Centralized resource allocation and target setting (CRATS) refers to a process in which top management has the power and responsibility to allocate limited input resources and to arrange expected output targets among its own decision-making units (DMUs). Data envelopment analysis (DEA) has been widely used to investigate the CRATS issues by considering the production process of DMUs as a single-stage structure, namely, a black box. This study proposes a novel centralized two-stage network DEA because the DMUs with a two-stage production process are a common phenomenon in real-world cases. The proposed model is applied to determine the optimal CRATS plan for an empirical case, where 27 retailers belong to an automobile parts supplier in Taiwan.