Optimal solutions of multiplier DEA models

Abstract Conventional models of data envelopment analysis (DEA) are based on the constant and variable returns-to-scale production technologies. Any optimal input output weights multiplier DEA these technologies interpreted as being most favorable for decision making unit (DMU) under assessment when latter is benchmarked against set all observed DMUs. In this paper we consider a very large class arbitrary polyhedral technologies, which includes almost known convex models. We highlight fact that conventional interpretation in such generally incorrect, raises question about meaning address prove solutions show DMU best light comparison to entire technology, but not necessarily This result allows clear meaningful models, including with complex constraint structure whose has been problematic left unaddressed existing literature.