AI-Driven Adaptive Learning Framework for Real-Time Optimization and Fault-Tolerant Control in Large-Scale Interconnected Computing Environments

Abstract

Modern large-scale interconnected computing environments require a paradigm shift toward autonomous, self-healing architectures to handle stochastic workloads and system failures. This paper proposes a multi-layered AI-driven adaptive learning framework for real-time resource optimization and Active Fault-Tolerant Control (A-FTC). Drawing from highly-cited recent advancements in deep reinforcement learning (DRL) and neural-based control, we integrate a hierarchical engine that balances operational efficiency with structural resilience. Experimental simulations demonstrate that our framework achieves high classification accuracy in fault isolation and a significant reduction in tail latency compared to traditional heuristic-based models.