A stochastic tri-level interdiction model for relief train location and infrastructure protection

• We introduce a strategic Defender-Attacker-Defender model for relief train deployment and infrastructure protection. • Our approach incorporates scenario-dependent disruptions at the attacker-stage to enhance robustness in disaster response planning. • We develop a progressive hedging-inspired algorithm to tackle the problem’s computational challenges. • Our results show that accounting for uncertainty and interactions can improve overall survival probability by up to 70%.