International Journal of Scientific Research and Engineering Development

Distributed computing environments have become the backbone of modern cloud infrastructure, high-performance computing (HPC) clusters, and large-scale data processing pipelines. The central challenge in such systems is efficient task scheduling — the process of allocating computational tasks to available nodes in a manner that maximizes throughput, minimizes latency, and balances workload distribution. Traditional scheduling algorithms such as Round Robin, First-Come-First-Served (FCFS), and static priority queues fail to adapt dynamically to fluctuating resource availability and heterogeneous task demands. This paper introduces SchedAI, an intelligent, AI-driven task scheduling framework that combines Reinforcement Learning (RL) with a hybrid heuristic-predictive engine to make real-time, context-aware scheduling decisions in distributed systems. SchedAI employs a Deep Q-Network (DQN) agent trained on historical workload traces, dynamic resource telemetry, and task dependency graphs to learn optimal scheduling policies without manual rule engineering. The system further incorporates a multi-objective optimization layer balancing energy efficiency, execution time, and fault tolerance simultaneously. Experimental evaluation on benchmark distributed workloads demonstrates a 34.7% reduction in average task completion time, 28.3% improvement in resource utilization, and 41.2% reduction in energy consumption compared to conventional baselines. SchedAI provides a deployable, scalable solution applicable across cloud platforms, edge computing environments, and enterprise data centers.

A.Agesta Jenifer, Dr.N.Aminur Rahman,"An AI-Based Intelligent Task Scheduling Framework for Distributed Computing Systems" International Journal of Scientific Research and Engineering Development, V9(3): Page(1764-1770) May-June 2026. ISSN: 2581-7175. www.ijsred.com. Published by Scientific and Academic Research Publishing.