Data envelopment analysis network computing in multi-stage parallel processes

DEA (Data Envelopment Analysis) is a managerial method that has been widely used for performance analysis in various public and private sectors. To deal with large-scale DEA problems, this research proposes the architecture of DEA network computing (LAN: Local Area Network) that is designed to coordinate a simultaneous use of multiple personal computers. An important feature of the proposed DEA computer architecture is that it is computationally structured in multistage parallel processes to enhance its algorithmic efficiency. The performance of the proposed network computing approach is tested and examined in a large simulation study.