Abstract
<title>Abstract</title> <p>Geopolitical disruptions can severely degrade supply chain performance by causing transportation restrictions, facility-capacity losses, procurement shocks, and sudden surges in demand. These effects are particularly critical in humanitarian and emergency logistics networks, where decisions must balance cost efficiency, service continuity, delivery responsiveness, and resilience under uncertainty. This paper proposes a preference-guided multi-objective optimisation framework for supply chain decision support under disruption. The problem is formulated as a scenario-based mixed-integer network optimisation model with four conflicting objectives: minimising expected total cost, un- met demand, delivery delay, and resilience loss across disruption scenarios. To solve the model, a preference-guided non-dominated sorting genetic algorithm II (PG-NSGA-II) is developed. The proposed method incorporates decision-maker preferences directly into the evolutionary search through an adaptive preference-pressure mechanism and a scenario-robust preference score, enabling the search to focus on decision-relevant regions of the Pareto front while preserving diversity. A multi- criteria decision-making layer based on TOPSIS and VIKOR is then used to rank the resulting non-dominated solutions and support actionable decision selection. Computational experiments on a disrupted humanitarian supply chain case study inspired by Middle East conflict conditions compare PG-NSGA-II with five benchmark algorithms: WSGA, MOPSO, SPEA2, MOEA/D, and classical NSGA-II. The results show that PG-NSGA-II achieves the best overall decision-support performance, reducing unmet demand, delivery delay, and resilience loss while improving demand satisfaction, preference alignment, and MCDM ranking quality. The proposed framework provides methodological and practical contributions for resilient supply chain optimisation in high-risk environments. Additional sensitivity, scalability, and reproducibility analyses are included to clarify the robustness of the proposed approach and its practical relevance for humanitarian logistics decision makers.</p>