A Decision Support Model for Transportation Companies to Examine Driver Behavior

Abstract

The entire society, and particularly the transportation companies, has an interest to improve traffic safety. Besides more than 3000 lost lives on the roads every day, there are significant financial consequences of road traffic accidents (RTAs). The purpose of this article is to design an efficient model for providing information about driver propensity for RTAs based on assessing their personality traits. This is achieved by creating a fuzzy inference system (FIS) where inputs are the scores from the implemented psychological instruments and output is the number of RTAs experienced by a driver. To adjust the functioning of FIS to the empirical data, a bee colony optimization (BCO) metaheuristic is applied. In this optimization procedure, we test three approaches for defining the variables of initial FIS and compare their performance. Simulation results show the differences between the considered approaches and, generally, very promising achievements of the proposed algorithm. The best-found FIS reached a 36% improvement of the objective function compared to the starting FIS. This FIS can be used, inter alia, as a decision-making tool in recruitment procedures for professional drivers to assess their propensity for RTAs, by that saving the lives of people and reducing the costs of the companies.