Structure of Data Envelopment Analysis with Intuitionistic Fuzzy Data (Case Study: Evaluation of Safety-Based Performance in Construction Projects)

Data Envelopment Analysis, one of the keys subfields in mathematical planning, has found a prominent role in evaluating the efficiency of decision making units in various fields.

In this research, after analysis and initial processing, based on studies, a parametric and efficient method for ranking intuitionistic fuzzy numbers is selected and proposed. The correctness of the performance of the selected method is obvious due to its formulation in linear structures. The developed model of data envelopment analysis, its mathematical formulation by CCR and IO-BCC methods are expressed in terms of governing the model structure and its implementation approach. A case study is presented to determine the factors affecting safety performance using the model. Based on previous theoretical studies and opinions of experts in the field of safety, the most important influencing factors (work pressure and perception of the supervisors' safety as inputs) and (the rate of physical and mental injuries and unsafe accidents as outputs) were selected. In addition to ranking the units, sensitivity analysis was performed in CCR and IO-BCC methods to rank the specified indicators in the inputs and outputs, and the results have been compared.

The results of the data envelopment analysis model with intuitionistic fuzzy data showed that with increasing k, the number of efficient units increases. On the other hand, in CCR and IO-BCC methods, the lowest and highest efficiencies belong to the pessimistic view (k = 0) and the balanced view (k = 0.5), respectively. Sensitivity analysis also showed that, in CCR and IO-BCC methods, the work pressure is the most safety factor affecting the efficiency results.

Originality/Value: 

Using a Data Envelopment Analysis model with intuitionistic fuzzy data to evaluate the performance of construction sites from a safety perspective can provide significantly better results. Because in the real world, there is uncertainty, and intuitionistic fuzzy data, due to the concept of belonging, non-belonging, and suspicion in the view of decision-makers simultaneously and in data reporting, is of particular importance.