Implementation of flexible DEA structure in the management of business processes

Abstract

This paper presents some statistical tests that can be used to test hypotheses of interest for implementing new technologies and improving the efficiency, functionality and business performance of business systems or companies. DEA is a non-parametric technique based on linear programming for measuring the efficiency of decision-making units (DMU) with heterogeneous inputs/outputs. Generally speaking, DEA has a possible large number of inputs and outputs, and the essential problem is that it is often not clear which inputs and which outputs should be chosen when applied to a real problem, that is, when evaluating the efficiency of the system. In tests, the deviation from the DEA borderline can be viewed as a stochastic variable. The DEA estimate is certainly biased in finite samples (debatable statistics), while the expected value of the DEA efficiency is almost certainly the true value of the parameter in large samples (complete statistics). The application of a flexible DEA structure is significant, if there are situations where insufficient information can prevent the use of parametric statistical tests in the processes and outcomes of management and production. The possibility of using DEA analysis in researching the efficiency of constituent units (branches; decision-making unit - DMU) of business systems should be taken into account in research concerning the functionality of business systems and their optimization in terms of productivity, efficiency and functionality. The research results in this paper are presented in tabular and graphical form.