Simulation Environment for Large-scale Evacuation : Modelling Approach and Experimentation

This report details the development of a comprehensive simulation environment designed to model and analyse large-scale evacuation processes. The primary objective is to provide a structured, operationally relevant framework that represents individual and group responses to evacuation orders, movement through transportation networks, and the influence of infrastructure and resource availability on overall evacuation performance. The environment integrates discrete-event simulation with intelligent agent modelling to operationalize established evacuation principles within a testable and verifiable framework. Development, verification, and experimentation processes are documented, emphasizing model robustness and operational relevance. A case study based on an evacuation scenario in Karlshamn demonstrates the simulation’s ability to capture behavioural dynamics and logistical constraints, enabling scenario analysis, performance assessment, and evidence-based regional preparedness planning. Results highlight strong sensitivity to background traffic, significantly affecting rescue times and overall system performance. In the baseline scenario, private car use achieves the highest efficiency despite infrastructure limitations, whereas public transport - including buses and trains - is constrained by capacity and scheduling dependencies, potentially reducing performance without coordinated implementation. Simulation outputs align with observed traffic dynamics, supporting model validity. The framework also captures the effects of checkpoint configurations, modal distribution, and multi-destination routing on evacuation outcomes.