Digital Twin Framework for Machine Learning-Enabled Integrated Production and Logistics Processes

This paper offers an integrated framework bridging production and logistics processes that employs a machine learning-enabled digital twin to ensure adaptive production scheduling and resilient supply chain operations. The digital-twin based architecture will enable manufacturers to proactively manage supply chain risk in an increasingly complex and dynamic environment. This integrated framework enables "sense-and-respond" capabilities, i.e. the ability to sense potential supplier and production risks that affect ultimate delivery to the customer, to update anticipated customer delivery dates, and recommend mitigating steps that minimize any anticipated disruption. In its core functionality this framework senses disruptions at a supplier facility that cascade down the upstream supply chain and employs the predictive capabilities of its machine learning-based engine to trigger and support adaptive changes to the manufacturer's MES system. Any changes to the production schedule that cannot be accommodated in a revised schedule are propagated across the downstream supply chain alerting end customers to any changes.