Historical data-driven risk assessment of railway dangerous goods transportation system: Comparisons between Entropy Weight Method and Scatter Degree Method.

In this paper, two historical data-driven weight calculation approaches including Entropy Weight Method (EWM) and Scatter Degree Method (SDM), are applied and compared to solve the risk assessment of railway dangerous goods transportation system (RDGTS). The risk is defined as product of occurrence probabilities and their corresponding weights, the occurrence probability and weight of each risk sub-indicator is obtained by using the historical statistical data as the inputs. EWM aims at reflecting relative intensities among the evaluation indicators, however, SDM reflects the projection factor that maximizes the overall difference among the evaluated objects. In order to compare the performance of EWM and SDM, the stability analysis is carried out, which focuses on studying and analyzing the reverse degree order of the weight value when the initial matrix changed. In addition, correlation coefficient between occurrence probability and final assessment results based on the two approaches are conducted. The results of case study by using China data show that: the SDM is more stable than EWM; sub-risk indicators with highest ranks belong to risk factors of Human category, which means the Human unsafe actions and behaviors are the most dangerous factors that influence the normal and safe operations of RDGTS.