A Bayesian decision network–based pre-disaster mitigation model for earthquake-induced cascading events to balance costs and benefits on a limited budget

Abstract

Cascading disasters induced by earthquakes amplify the severity of the initial impact on environment. Although decisionmakers may face uncertainty, an effective mitigation strategy is critical in environmental management. We propose an earthquake-induced cascading disaster mitigation–Bayesian decision network (ECDM-BDN) model to assess pre-disaster mitigating strategies under limited budgets from the perspective of systematic thinking. This model graphically represents the complex relationship among various variables in a seismic hazard system and resistance system based on disaster system theory. It can predict the triggering of cascading events through probabilistic reasoning and identify the key variable accountable for the range of outputs observed through sensitivity analysis. In addition, cost–benefit analyses are carried out by combining Bayesian decision network utility nodes and dynamic programming to obtain a balance between costs and benefits in the context of limited budgets. An earthquake-induced liquefaction served as a case study to demonstrate the proposed model’s effectiveness. Experimental results indicate that the ECDM-BDN model can balance the costs and effects of each pre-disaster mitigating strategy as well as select the optimal one according to the utility value. The proposed model can perform a “white-box” decision-making process, which is expected to guide earthquake-induced cascading event pre-disaster mitigation in cases of limited budgets.