AI-Driven Autonomous Resource Scheduling in Cloud–Edge Environments

Abstract

Cloud Edge computing provides services with low latencies by allocating workloads between centralized cloud servers and geographically close edge nodes, but dynamic workloads and nonhomogeneous resources render the scheduling of workloads a thorny multi-objective optimization problem. The given paper proposes a framework of Autonomous Resource Scheduling (AIARS) which makes use of deep reinforcement in order to execute adaptive and real-time task placement across Cloud-Edge infrastructures. The scheduler models provide a model of system dynamics using state representations that include resource utilization, queue status, and task characteristics and learns policies that combine to reduce response latency, operational cost, and load imbalance. Scheduling is a reward-based decision process that is mathematically formalized to allow the continuous improvement of a policy depending on traffic conditions. The suggested structure is tested on the basis of the large-scale simulations of the workload pattern, and it is compared to the conventional heuristics such as Round Robin, Greedy, and First-Come First-Serve scheduling. Findings indicate that AIARS can reduce average response time by up to 25 percent and use considerably more resources without reducing convergence and scalability to bursty demand. Such results suggest that smart, self-adaptive scheduling can contribute significantly to the level of efficiency and robustness in distributed computing systems to support the next-generation latency-sensitive applications.