Cost minimizing target setting heuristics for making inefficient decision-making units efficient

Abstract

A problem of increasing the efficiencies of inefficient decision-making units (DMUs) with known marginal costs of reducing inputs is considered. It is shown that the problem is difficult to solve exactly, and has exponential computational complexity. Two heuristics (Heuristic-I and Heuristic-II) that consider cost minimizing DEA frontier projections for inefficient DMUs are proposed. The proposed heuristics are tested by using several simulated and real-world datasets with linear and Fibonacci marginal input reduction costs. The results of the proposed heuristics are compared with traditional cost insensitive DEA frontier projections, and it is shown that Heuristic-I results in approximately 18–58% lower costs.