Agent-based simulation optimization for dynamic disaster relief distribution

This work presents an agent-based simulation optimization framework to model the impact of transport disruptions and word of mouth on disaster relief distribution. An agent-based simulation considers uncertainties in transport conditions and further incorporates various actions by and interactions between multiple people affected by a natural hazard event. To select optimal distribution points and vehicle types, a bi-objective optimization procedure is implemented focusing on the minimization of costs and maximization of services provided. The developed solution procedure is tested on a sample setting based on the April 2015 earthquake in central Nepal. Computational experiments study the impact of transport disruptions and word of mouth on disaster relief planning. Results indicate the importance of considering such factors in planning procedures.