Dynamics and risk assessment of a remanufacturing closed-loop supply chain system using the internet of things and neural network approach

To address the problem of the large subjective error of expert evaluation methods in supply chain management, the supply chain system is comprehensively analyzed, and a deep learning backpropagation (BP) neural network-based supply chain risk assessment model is constructed. First, the basic theories of supply chain and risk assessment are described, and the process of supply chain risk management is explained. Then, the ANN (artificial neural network) is discussed in detail. On this basis, the feasibility of the BP neural network applied in the risk assessment of the supply chain is analyzed. In addition, the risks of the supply chain system are analyzed under the support of the Internet of Things (IoT), and the indices for risk assessment of the supply chain are determined. The reliability analysis, validity analysis, and factor analysis of the evaluation indices are implemented using a questionnaire survey, based on which the risk assessment indices of the supply chain are determined as 7 first-level indices and 20 sesond-level indices. Finally, a BP neural network-based supply chain risk assessment model is established, and the simulation results are analyzed in MATLAB. The maximum relative error of the proposed BP neural network model for supply chain risk assessment is as low as 0.03076923%, and that calculated by the AHP (analytic hierarchy process) is 57.41%. Compared with that of AHP, the fitting degree of the BP neural network-based supply chain risk assessment model is much higher. Meanwhile, the simulation experiment indicates that the established risk assessment model has strong generalization ability and learning ability. This work not only provides technical support for the development of remanufacturing closed-loop supply chain systems but also contributes to the improvement of the accuracy of supply chain risk assessment.