Cross efficiency based heuristics to rank decision making units in data envelopment analysis

- Propose cross efficiency heuristics (CEHs) as an alternative for traditional CEs.
- Propose a stratification-based consistency evaluation framework (SCEF).
- No Linear Programming used in CEHs. Thus, no multiple optimal weight problem occurs.
- SCEF used to evaluate performance of CEHs and other CE-based ranking methods.
- Numerical examples demonstrate that CEHs are comparable to CE-based ranking methods.

This paper proposes two cross-efficiency based heuristics (CEHs) for ranking decision making units (DMUs) in Data Envelopment Analysis (DEA). This paper also suggests a stratification-based consistency evaluation framework (SCEF) to evaluate the performance of the proposed heuristics and several DEA-based full ranking methods in terms of consistency. Several cross-efficiency (CE) methods have been developed as a DEA extension to rank efficient and inefficient DMUs with the main idea of using DEA to do peer evaluation. However, it has been well known that they all suffer from lack of discrimination since efficiency scores from those methods may not be unique due to the non-uniqueness of the DEA optimal weights in the Linear Programming (LP) models. The proposed CEH methods overcome this issue since they do not use any LP models but show comparable consistency level to other DEA-based full ranking methods based on SCEF. Numerical examples from the literature and simulation indicate that CEHs are a good surrogate for CE method.