Complex System Risk Evolution Analysis Based on Multi-resolution Modeling and Bayesian Networks

Complex system development contains processes full of comprehensive integration, complicated iteration and highly coupling, and ultimately come into being hierarchical networks of risk evolution. This paper presents a new approach on risk evolution analysis by multi-resolution modeling describing complex system and Bayesian network learning risk relation networks. By simulation experiments, risk evolution networks and their comparative analysis on different resolution models are presented. The proposed approach is verified to provide the more credible assessment and avoid risk evolution networks just only determined by subjective judgment.