A modification of a mixed integer linear programming (MILP) model to avoid the computational complexity


ÖRKCÜ H. H., ÜNSAL M. G., BAL H.

ANNALS OF OPERATIONS RESEARCH, cilt.235, sa.1, ss.599-623, 2015 (SCI-Expanded) identifier identifier

  • Yayın Türü: Makale / Tam Makale
  • Cilt numarası: 235 Sayı: 1
  • Basım Tarihi: 2015
  • Doi Numarası: 10.1007/s10479-015-1916-3
  • Dergi Adı: ANNALS OF OPERATIONS RESEARCH
  • Derginin Tarandığı İndeksler: Science Citation Index Expanded (SCI-EXPANDED), Scopus
  • Sayfa Sayıları: ss.599-623
  • Anahtar Kelimeler: Data envelopment analysis, Discriminant analysis, Suitable input-output weights, Cross efficiency, Ranking, DATA ENVELOPMENT ANALYSIS, CROSS-EFFICIENCY EVALUATION, EUROPEAN ELECTRICITY DISTRIBUTION, PERFORMANCE EVALUATION, INTERNATIONAL BENCHMARKING, DEA, WEIGHTS, PRODUCTIVITY, OPERATORS, COST
  • Gazi Üniversitesi Adresli: Evet

Özet

Having multiple optimal solutions to weights affects to a great extent the consistency of operations related to weights. The cross efficiency method is the most frequently studied topic in data envelopment analysis (DEA) literature. Originally, the cross efficiency method included the efficiency evaluations that were obtained for a decision making unit (DMU) by the classical DEA for the reuse of optimal weights in other DMUs. As the optimal weights in classical DEA solutions usually have multiple solutions, this reduces the usefulness of the cross evaluation. Lam (J Oper Res Soc 61:134-143, 2010) proposed a mixed-integer linear programming (MILP) formulation based on linear discriminant analysis and super efficiency method to choose suitable weight sets to be used in cross efficiency evaluation. In this study, Lam's MILP model has been modified to reduce the steps during the solution process. The model also becomes a linear programming model after the modification to make it easier to use and to reduce the computational complexity. Numerical examples indicate that the proposed weight determination model both reduces the steps and minimizes computational complexity. Furthermore, it has similar performance with Lam's MILP model for the cross efficiency evaluation.