Stochastic optimization for investment in facilities in emergency prevention

• Three measures are considered synthetically for earthquake prevention.
• A two-stage stochastic programming model is formulated.
• CVaR is employed to optimize the tradeoff between medium- and worst-case scenarios.
• An earthquake in Lushan County of Sichuan, China is examined.

A coordinated approach is developed to integrate three preventive measures (i.e. building reinforcement, reinforcement of road networks, and facility location of relief supplies), with the objectives of minimizing budgets and risk-induced penalties. The Conditional Value-at-Risk is employed as a decision-making tool to evaluate diverse decisions of prevention based on the degree of risk aversion. Based on a real-world case of an earthquake, a series of scenarios were designed, and the applicability of the proposed model was studied. The coordinated approach for investing preventive measures is cost-efficient in helping reduce the impact of disaster on society.