Environmental Assessment on Energy and Sustainability by Data Envelopment Analysis

10NETWORK COMPUTING

10.1 INTRODUCTION

This chapter¹ proposes a computing architecture that is computationally structured by multi‐stage parallel processes to enhance its algorithmic efficiency in solving various DEA problems. Most of previous computations were traditionally performed by a computer with a single processor without much thought for computational efficiency and time savings. The computational approach proposed in this chapter may not produce a computational benefit for small DEA problems or one‐time applications. However, when a problem becomes very large (e.g., more than a few thousand DMUs), the traditional approach for computation becomes painfully insufficient and very time‐consuming². The proposed computational scheme is useful in dealing with such large DEA problems, including DEA environmental assessment.

The purpose of this chapter is to document how network computing technology can be linked to DEA performance evaluation on spreadsheet software. It is expected that the proposed network computing can solve large problems with a considerably reduced central processing unit (CPU) time. It is true that we do not have any question on the modern use of personal computers (PCs) for DEA. Their technology development is amazing and very effective. However, it is important for us to discuss how to connect multiple PCs for handling a large problem such as a bootstrap approach that needs many re‐sampling processes in order to outline the shape of an efficiency frontier and ...