Answers are grounded in this paper and cite their sources.
Cloud computing environments face significant challenges in maintaining optimal performance due to dynamic workloads and heterogeneous resource configurations. Traditional load balancing algorithms often fail to adapt to real-time fluctuations, leading to suboptimal resource utilization, increased latency, and degraded performance. This paper presents a novel hybrid load balancing model that integrates Long Short-Term Memory (LSTM) networks with Ant Colony Optimization (ACO) and Throttle algorithms. The proposed approach leverages LSTM's predictive capabilities to forecast workload patterns, ACO's metaheuristic optimization for intelligent task scheduling, and Throttle mechanisms for dynamic resource allocation. Experimental evaluation using CloudSim-based simulation demonstrates that the hybrid model achieves a response time of 120 ms, throughput of 150 tasks/second, and resource utilization of 85%, significantly outperforming configurations lacking individual components. The LSTM model achieved a Mean Absolute Percentage Error of 2.17% and inference time of 0.7249 seconds, confirming its suitability for real-time applications. Ablation studies validate the synergistic contribution of each component, with the complete hybrid model consistently outperforming ablated versions across all performance metrics.
No takes yet. Share an insight, caveat, or question.
Michael Adeniyi Ibiyomi (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: