Distributed Dynamic Offloading Flexible Strategy of Tasks in Digital Converter Stations for Industrial Internet of Things

The rapid development of smart grids has created new challenges for digital converter stations, which are essential components of smart grids and must address the need for distributed computing and task offloading to process the massive data generated by the Industrial Internet of Things (IIoT), thereby driving the demand for flexible manufacturing.The dynamic adjustment of system production tasks and resource scheduling is both highly complex and time-sensitive. Traditional centralized scheduling methods suffer from computational bottlenecks and high energy consumption in multi-user scenarios, making it difficult to meet the real-time processing demands of complex tasks.To address these challenges, this paper proposes a two-layer edge-converged terminal network architecture, which simplifies system modeling, reduces energy consumption, lowers costs, and is easy to monitor and manage. Furthermore, leveraging the concept of flexible manufacturing, we introduce a distributed dynamic computation offloading model based on deep reinforcement learning. By incorporating the Deep Deterministic Policy Gradient (DDPG) algorithm, the model enables adaptive task offloading and computational resource allocation, reducing reliance on central scheduling while meeting the flexible requirements for dynamic task scheduling and resource optimization.The model’s performance under various task loads and energy efficiency requirements is evaluated through simulation tests and compared with other classical offloading algorithms. The results show that the proposed strategy outperforms traditional methods in terms of task processing delay and energy management, demonstrating superior convergence and stability. Additionally, through a dynamic weight adjustment mechanism, the model effectively balances delay and energy efficiency under different load scenarios in the converter station, significantly improving the operational efficiency and resource scheduling flexibility of the converter station in IIoT-based flexible manufacturing.