Autonomous pod-based railway systems represent a promising innovation in freight transport, combining the flexibility of modular vehicle concepts with the efficiency of rail-based logistics. Their success, however, depends on effectively managing the assignment and relocation of carriers under dynamic and uncertain operating conditions. Static optimization approaches such as Mixed-Integer Linear Programming (MILP) can design efficient baseline schedules, but these often prove fragile when confronted with real-world uncertainties such as delays, carrier breakdowns, and structural disruptions. To address this gap, this thesis develops a hybrid decision-making framework that integrates MILP-based planning with a Discrete-Event Simulation (DES) environment, enabling dynamic re-optimization and real-time resilience analysis.

The framework operates through an event-driven feedback loop: initial assignments are optimised using MILP, executed within the DES, and re-optimised whenever disruptions occur. This coupling allows for continuous adjustment of plans to reflect the real-time state of the system. The methodology was validated through two case studies: a simplified Toy Case to verify the model logic and a large-scale Randstad network to evaluate system robustness against cascading disruptions. The analysis incorporated a range of scenarios, including probabilistic delays, deterministic breakdowns, arc removals, transport unit (TU) insertions at varying time brackets, and the impact of delivery-window flexibility.

Results demonstrate that re-optimisation is highly effective in recovering service levels after disruptions, significantly improving fulfillment rates and resource utilisation compared to static schedules. The system showed strong adaptability to sudden demand surges and carrier failures by reallocating idle resources, while also highlighting resilience thresholds in cases of severe network degradation. Time-window flexibility was found to play a dual role: it improved overall fulfillment but introduced delays, suggesting trade-offs that need to be balanced by operators and policymakers. Carrier utilisation and empty travel metrics further revealed how resilience is achieved at the cost of increased repositioning.

This study contributes both theoretically and practically. It establishes an integrated simulation–optimisation framework that advances research on disruption management in autonomous rail systems, and it provides operational insights on fleet pre-positioning, flexible service design, and the role of redundancy in network infrastructure. The findings emphasise that digital re-planning capabilities are necessary for future autonomous freight systems, while industrial and policy implications include incentivising early bookings, setting flexibility standards, and ensuring investment in redundant routing capacity. Overall, the proposed framework provides a powerful tool for designing, testing, and improving resilient autonomous pod-based rail networks, bridging the gap between strategic planning and operational execution under uncertainty.