Simulation optimization of supply chain uncertainty factors based on support vector machines optimized with improved Jellyfish Search Algorithm

To address uncertainties in the supply chain, this paper analyzes potential uncertainties by constructing a three-tier supply chain simulation model and proposes an optimized support vector machine algorithm (IJS-SVM) using the improved Jellyfish Search Algorithm to optimize these uncertainties. In comparison to the classical Jellyfish Search Algorithm, the proposed algorithm incorporates Logistic chaotic mapping, Cauchy variation, and inverse learning strategies, with the goal of improving solution efficiency and avoiding local optima. The experimental results indicate that IJS-SVM outperforms the support vector machine optimized by the classical Jellyfish Search Algorithm (JS-SVM), the support vector machine optimized by the genetic algorithm (GA-SVM), the support vector machine optimized by the particle swarm optimization algorithm (PSO-SVM), and the traditional support vector machine (SVM) in terms of classification accuracy. Building on this study, the SHAP explanatory model is employed to interpret IJS-SVM’s prediction results, quantify the specific impact of uncertainty factors on each stage of the supply chain, and input these quantitative results into the simulation system to assess warehouse allocation and identify strategies to enhance warehouse utilization, ultimately minimizing supply chain costs. The experiment demonstrates that the model not only offers valuable practical guidance for CEPC engineering project planning but also has broad applicability and can be extended to uncertainty management and optimization within the tertiary supply chain sector.