The Multi-Objective Optimization Algorithm Based on Monte Carlo Simulation in Supply Chains

Abstract

With the increasing complexity of supply chain management, how to effectively balance multiple objectives, such as cost, efficiency, and service level, has become a key challenge in optimization problems. To address the issue of insufficient optimization in complex supply chain scenarios, this study proposes an improved multi-objective optimization algorithm (MOA) based on Monte Carlo simulation (MCS). The algorithm utilizes a support vector machine (SVM) to establish a surrogate model, enhancing simulation efficiency, and combines MCS with a genetic algorithm (GA). For the procurement cost budget of large retail company A in 2020, the proposed method predicts the mean square error (MSE) for the four quarters to be between −1.3 billion and 2.2 billion, with a uniform and stable distribution across quarters. Moreover, regardless of whether the cost budget increases or decreases, the number of product defects per million remains below 1370, and the predicted product delivery delay can be as short as approximately six days. After 10 training iterations, the algorithm optimizes the delivery time to two days. The results indicate that the proposed MOA can effectively improve the solution efficiency and optimization performance of multi-objective supply chain problems, providing an efficient and stable tool for supply chain optimization. This method can be applied to complex supply chain issues in real-world scenarios and also offers new insights for the improvement of MOAs.