Research on Intelligent Factory-oriented Edge Cloud Collaborative Automated Scheduling and Resource Optimal Allocation Technology

Smart factories are faced with the challenges of heterogeneous equipment, dynamic tasks and bandwidth and delay of massive data processing. Traditional centralized cloud computing scheduling has high delay, bandwidth pressure and single point failure risk, while single edge computing is limited by insufficient resources. This study focuses on the automatic scheduling and resource optimization allocation technology under the edge cloud collaborative architecture. This paper designs a three-tier edge cloud collaboration architecture of "hierarchical awareness-dynamic collaboration", and realizes refined resource management and collaboration through edge layer resource awareness agent, collaborative layer dynamic service bus and cloud layer digital twin engine. A hybrid scheduling algorithm named "Dynamic Divide and Reinforcement Scheduling" (DDRS) is proposed, which combines the dynamic divide and conquer of tasks with feedback reinforcement learning mechanism to achieve low delay and high robustness scheduling. A multi-dimensional game-cooperative equilibrium resource allocation model is constructed, and multi-resource joint optimization is realized based on Nash equilibrium and ADMM (Alternative Direction Method of Multipliers) algorithm, and an elastic migration mechanism is designed to deal with resource overload. The experimental results show that compared with pure edge and pure cloud scheduling, DDRS reduces the average delay to 63ms and the overtime task rate to 5.3%. The resource optimization model reduces the standard deviation of resource utilization from 0.41 to 0.18, reduces the number of elastic migration to 9 times, and can recover within 120ms when the node fails, which verifies the effectiveness of the technology in improving efficiency, stability and fault tolerance. This research provides theoretical and practical support for the intelligent upgrading of the whole chain of smart factories.