Blockchain acceptance rate prediction in the resilient supply chain with hybrid system dynamics and machine learning approach

Abstract

In today’s era, the importance and implementation of blockchain networks have become feasible as it improves the resilience of the supply chain network at all levels by clarifying information and creating security in the network, improving the speed of response, and gaining the trust of customers. This paper aims to investigate the behavior of the blockchain acceptance rate (BAR) in the home appliances flexible supply chain in Iran using. system dynamics (SD), which is used to better define the relationships between the variables of the model that are non-linearly connected. Through simulating the behavior of the BAR in the long term in the supply chain, whilst conducting sensitivity analysis, policy design, and validation, this model will be implemented for the years 2020 to 2030. Additionally, post-simulation, blockchain acceptance behavior will be assessed by having simulated data considered as input for studied Multi-Layer Perceptron (MLP) and Vector Regression (SVR) (data that have the highest correlation with BAR). The acceptance rate behavior is predicted with the help of machine learning methods to have the best behavior and prediction for the data of 2020-2022 since the prediction function is compared to daily real data obtained these years. The results show that in 2030, the BAR will be around 0.6 if the COVID-19 outbreak impact is medium, and if the considered policy designs are implemented, this rate will reach a maximum of 0.8. So paying attention to the creation and design of policies can achieve positive implications for increasing the resilience of the supply chain in the long run. Findings suggest that the SD-MLP method is better than the SD-SVR method as it has less error and can predict the better behavior of the BAR.