Optimization design of supply chain network based on BP neural network performance evaluation and feedback mechanism

Abstract

SummaryThis paper proposes a supply chain network design method suitable for multi‐product and multi‐inventory models, and uses the improved BP neural network to evaluate and provide feedback on the collaborative performance of the supply chain, adjusting the supply chain network design scheme on time. In the context of the Internet of Things (IoT) in manufacturing, it has been found that supply chain operations are difficult to meet personalized customer needs with high precision and quality. Therefore, we adopted a dynamic library strategy, supply chain network optimization model, hybrid algorithm, and the improved BP neural network to solve the above problems. First, this paper designs a corresponding inventory strategy selection mechanism for the various ordering methods of retailers in the manufacturing IoT environment. Based on this, we have constructed a dual objective model for a sustainable supply chain network to minimize total cost and maximize customer satisfaction. Second, we have developed a hybrid improved Grey Wolf and Whale Algorithm (OLDGWOA) that can accurately solve the above model. The hybrid algorithm divides the population into two parts through opposition‐based learning, and then we use the improved grey wolf algorithm and whale algorithm to solve the two populations, and seek the optimal solution in the results, resulting in a hybrid algorithm. Finally, we constructed a supply chain performance evaluation model and feedback mechanism based on the improved BP neural network to adjust inventory strategies and network design at any time. We also validated the developed model and algorithm through numerical examples, and the results showed that: (1) the hybrid algorithm has certain advantages in search and solution speed, (2) the advantages of supply chain network design based on supply chain performance evaluation and feedback mechanisms, and (3) the trade‐off between ordering methods and inventory strategies, as well as the trade‐off between location and inventory strategies.