Robust human evacuation planning under hybrid uncertainty for large passenger ships

Recent updates to evacuation protocols have highlighted the necessity for robust models on passenger ships. They should refine the evacuation process, particularly by considering the detailed behaviors of passengers (e.g., alertness level) and addressing uncertainty in walking speeds among individuals and family groups. Additionally, these models should optimize the positioning of exit doors and factor in potential disruptions to their operation. This approach involves directing passengers to the most suitable exit doors, followed by the strategic assignment of crew members to assist passengers. This paper proposes an optimization model that aligns with updates and generates personalized evacuation plans. It minimizes evacuation time while optimizing the location of exit doors and crew members to reduce their distance from passengers. These objectives are constrained by factors like passengers' starting locations and the corridor's traffic flow. The model also accommodates the dynamics of families by introducing the family speeds. Furthermore, the model considers dynamic adjustments to evolving situations for walking speeds using exponential and linear functions. Then, a scenario-based stochastic technique is combined with a feasibility-based ρ -robust approach to manage uncertainty in passenger walking speed and disruption in exit door capacity. Finally, the model is tested through multiple test problems and further validated by a real-life Ro-Pax case study.