The Optimization of Load Balancing Strategies in Heterogeneous Distributed Environments for Regional Industrial Upgrading

INTRODUCTION: With the ongoing advancement of regional industrial upgrading, efficient load balancing in heterogeneous distributed environments has emerged as a pivotal technical challenge. Conventional load balancing methods, typically designed for homogeneous systems or based on static scheduling rules, are ill-equipped to handle the dynamic task demands and diverse resource capabilities characteristic of modern industrial computing infrastructures. OBJECTIVES: This study seeks to bridge this gap by developing an intelligent, adaptive load balancing strategy specifically tailored for heterogeneous environments, with the goal of maximizing resource utilization, minimizing response latency, and supporting sustainable industrial transformation. METHODS: The proposed framework synergistically combines real-time dynamic resource monitoring, multi-dimensional task characterization and scheduling, and reinforcement learning-driven decision-making to construct a responsive and self-optimizing load distribution mechanism that continuously adapts to system state changes. RESULTS: Comprehensive experiments show that the approach significantly outperforms traditional strategies (e.g., round-robin, shortest job first) and state-of-the-art deep reinforcement learning methods across key performance metrics (p < 0.05), achieving up to a 44% reduction in average response time and a remarkable 91.4% system-wide resource utilization. CONCLUSION: By enhancing both adaptability and efficiency in complex heterogeneous settings, the proposed strategy offers a practical and scalable solution with strong potential for deployment in smart manufacturing, cloud computing, edge computing, and other industrial digitalization scenarios.