Digital Twin-Enabled Adaptive Production and Asset Coordination for Resilient Industrial Operations Under Dynamic Disruptions

Abstract

Industrial production systems face disruptions caused by machine failures, material delays, workforce shortages, and changing production requirements. Conventional scheduling and maintenance practices often operate separately, limiting coordinated responses to changing operational conditions. This study proposes a Digital Twin enabled adaptive production and asset coordination framework that integrates machine degradation, Remaining Useful Life (RUL), workforce capacity, material availability, and disruption states within a virtual production environment. A co-simulation layer represents machine workforce interactions, while an adaptive multi-objective scheduling layer considers makespan, schedule deviation, energy consumption, material waste, recovery time, and operating cost. An Adaptive Twin-Reinforcement Learning (ATRL) layer incorporates carbon and waste considerations into disruption recovery decisions. A hierarchical edge cloud architecture supports local schedule repair and cloud-based policy updates. The framework is evaluated through synthetic simulation scenarios involving machine failures, demand surges, supply delays, workforce shortages, and simultaneous resource disruptions. Results show that the proposed framework reduces makespan from 120 h to 94 h and recovery time from 34 h to 17 h compared with static scheduling. Lower energy consumption and material waste are also observed across the evaluated scenarios. The findings indicate that Digital Twin-based adaptive coordination can support resilient production planning, asset management, sustainability-aware recovery, and operational decision-making under dynamic industrial conditions.