Hybrid Belief Rule Base Inference Method And Its Application In Safety Assessment Of Complex Systems

It is of great importance to conduct safety assessment of a complex system to ensure the safe operation of the system as well as to avoid accidents and loss of life and property.It needs to deal with different types of uncertain information and the non-linear relationships among multiple factors.Moreover,the constructed model is also required to be with a high level of interpretability.The belief rule base(BRB)can be used to represent various uncertain information and model nonlinear complex systems.Furthermore,it’s modeling and reasoning process is highly interpretable.The current BRB is constructed under a single assumption,e.g.,conjunctive or disjunctive.However,the practical conditions is much more complicated where the conjunctive and disjunctive assumptions coexist.The hybrid belief rule base is designed to deal with this situation.It follows the principle that the model must remain consistent with the practical conditions with a higher level accuracy so that the requirements of the safety assessment of a complex system can be met.This paper carries out research on hybrid belief rule base,its modeling and reasoning methods,and the application in the safety assessment of a complex system.The main contents are as follows:(1)The modeling and reasoning of hybrid belief rule base.The basic concepts of the hybrid belief rule base are proposed,including the definition of attribute combination,hybrid belief rule,and hybrid belief rule base.Further analysis of the relationship between the hybrid belief rules and the conjunctive and disjunctive belief rules.On this basis,a modeling and reasoning method of hybrid belief rule base is proposed.(2)A railway transportation safety assessment method based on hybrid belief rule base is proposed.Five aspects are summarized and analyzed,including environment,equipment,management,carrying capacity,accident prevention and rescue capabilities.The characteristics of the railway transportation safety assessment and the feasibility of using the hybrid belief rule base are analyzed,and a hybrid belief rule base model for safety assessment is proposed.(3)Safety assessment of railway transportation in Chen-Yu region is conducted using the proposed safety assessment approach using hybrid belief rule base.According to the characteristics of the high-speed line and existing line in this region,a hybrid belief rule base model is constructed.The effectiveness of the proposed model is verified through a comparative analysis with historical accident and fault reports.Upon comparing the safety assessment results of high-speed line and existing line,it is concluded that the high-speed line is safer than the existing line,and environmental and equipment factors have a greater impact on security.Moreover,it is found that the railway safety reports of Canada and Japan are highly consistent with what is concluded in this study.On this basis,further suggestions for improving the safety of railway transportation are put forward.