An agent-based simulation of post-disaster relief and medical assistance activities

Disaster management is commonly defined by four phases as planning, mitigation, response and recovery, of which the response phase is a critical yet challenging one to minimize fatalities under tight time and scarce resource constraints. Especially for the post-disaster medical assistance activities, how to optimize the rescue team dispatch based on their specialties and the damaged condition, and how to optimize the allocation of injured patients to hospitals based on their symptoms and hospital abilities, considering hospital capacities and transportation distance to minimize the fatalities, are forever crucial issues. Rather than handling above-mentioned problems separately, we aim to tackle them as part of an agent-based framework by which the resource allocation and subsequent inlocation resource scheduling are treated as a holistic system. This work only focuses on the two dispatching and allocation phases of which we apply real-coded genetic algorithm to minimize the total time cost of transportation in terms of distance and the disparity between the demand and supply of resources. The scenario analysis based on the simulated results could be used for post-disaster medical assistant training purposes and provide insights for policy-makers in future work.