Probabilistic modeling of integrated school-road systems for multi-hazard resilience

Abstract

Performance assessment of education and associated infrastructure is an integral part of disaster risk reduction of communities from natural hazards such as earthquakes and floods. Depending on the extent of structural or functional damage to buildings and accessibility to school compounds through road networks, the school compounds could be immediately usable, partially functional, or totally unusable after a hazard event. Some of the schools might also be used as shelters or evacuation centers, inevitably causing disruption to the education process. In order to model the disruption and develop strategies to minimize the disruption, a system-level probabilistic assessment is proposed in this paper, using a combination of Agent-based (AB) and Bayesian network (BN) approaches.

The BN component estimates the disruption to education due to multiple hazards by modeling causal effects and interaction between different interacting factors from physical, functional, and social vulnerability aspects of the infrastructure. The AB component explores strategies for recovery plans by detailed modeling of the road and bridge network, connecting the schools to the communities and shelters. This component considers decisions taken by two agents, namely the school and road operators, and explores the interaction between them. The paper presents preliminary findings of this study being developed under the UNESCO Chair on Disaster Risk Reduction and Resilience Engineering at UCL, based on a case study in Cagayan De Oro in the Philippines.