QUALITY-DRIVEN DEA-BASED DECISION SUPPORT SYSTEM: APPLICATION TO A CHINESE BATH & KITCHEN COMPANY

Abstract

Data envelopment analysis (DEA) is one of the most representative methods for efficiency evaluation. However, despite the maturity of the DEA studies on models and variables selection algorithms, users of DEA have very limited guidance during the whole analysis process. Three obvious deficiencies have been identified. First is the lack of guidance on DEA model selection. Second is the lack of decision supports to determine the variables reduction algorithm. Thirdly is the deficiency of the DEA quality evaluation methodology. In this thesis, we provide a systematical quality-driven DEA-based decision support system for DEA users to follow. A measurement methodology based on the MCDM is proposed to evaluate the overall quality of DEA. The support system has been applied to the actual selection project of a Chinese bath & kitchen company and proved to be effective in solving practical problems.