Requirements regarding the modeling of human error from an evolutionary and multifaceted perspective

The triggering element of this research work consisted in the descriptive-analytical study carried out from the available specialized literature regarding the foundation and practical applicability of the current methods of human reliability analysis: the quantitative ones from the first generation, the qualitative ones from the second generation as well as the most new nonlinear dynamic methods, in correlation with recent improvements in individual phases of human reliability analysis approaches. On this basis, the authors propose the transition to a new acceptance/understanding/vision of the concept of human error, adapted to the current complex socio-technical systems, characterized by rapid and often unpredictable evolution. The need to develop methods for evaluating human reliability based on modeling and the use of tools such as Bayesian networks, Petri nets and/or fuzzy logic was argued, while also pointing out the basic requirements that must be met. Ways to approach occupational and industrial safety challenges adapted to the rigors imposed by the multiple facets of the challenges highlighted in terms of the human factor are also proposed.