A bi-objective optimization approach for multi-echelon supply network disruption risk assessment and critical supply path identification

This study tackles the critical challenge of managing uncertainty and disruption risk in multi-echelon supply networks (SNs) under data scarcity. We propose a novel bi-objective optimization approach integrating Bayesian Networks (BNs) and information theory to simultaneously assess disruption risk and identify the critical supply path. To address data scarcity and inherent ambiguity, BN parameters (probabilities) are represented as interval values. Entropy and information gain are used to quantify uncertainty and the value of supplier information. Two nonlinear optimization models are developed to find pareto optimal solutions representing the trade-off between minimizing the downstream manufacturer’s fully disrupted probability (reflecting operational continuity concerns) and maximizing the total information gain obtainable from supply paths (guiding uncertainty reduction efforts). The models explicitly consider two crucial scenarios based on whether the manufacturer initially possesses information about none or some of its direct suppliers. We employ a problem-specific Non-dominated Sorting Genetic Algorithm II (NSGA-II) to solve the models. Experimental results on various SN topologies demonstrate the framework’s effectiveness. Findings indicate that identifying the critical supply path via information gain provides valuable insights for anticipating risks and strategically reducing uncertainty. Importantly, the determination of the critical supply path and the nature of the trade-off between risk and informational advantage are shown to be significantly influenced by both the SN’s topological structure and the manufacturer’s initial information state. This research offers a comprehensive decision support tool for enhancing SN resilience under uncertainty.