Deep Reinforcement Learning for Dynamic Compliance Management in Blockchain Supply Chains: An LSTM Perspective

Supply chain management is one of the industries that benefit greatly from implementing blockchain technology since it supplies chains with a high level of transparency, security, and traceability. Nonetheless, keeping up to date with different regulatory requirements or specifications that change frequently and still meeting contract requirements still pose a major problem. In this paper, we introduce a new method in the dynamic compliance management of blockchain implemented supply chain system by the use of deep reinforcement learning (DRL) and Long Short-Term Memory (LSTM) networks. Through DRL, supply chains can obtain real-time insights in learning from the environment and apply the knowledge effectively to satisfy regulatory standards as well as contractual terms. The inclusion of the LSTM component allows the system to capture sequential dependency and long-horizon effects within the supply-chain data. To prove the efficiency of our strategy, we provide simulation results as well as the analysis of real-world applications that discuss the positive shift in percentages and potential increase in productivity.