Scholarlink

<p>Energy efficiency remains a significant challenge in cloud computing, particularly due to the geographically distributed nature of cloud resources. Virtual Machine (VM) consolidation has emerged as a key strategy for optimizing resource utilization and reducing energy consumption. This paper presents a hybrid approach that combines Reinforcement Learning (RL), swarm-based metaheuristics, and the Modified Best Fit Decreasing (MBFD) scheduling algorithm to address energy-aware VM allocation and consolidation in cloud data centers. The proposed method integrates clustering techniques such as K-means and Cosine Similarity (CS) to enhance workload classification and decision-making. By focusing on software-based solutions and leveraging adaptive learning agents, the proposed method aims to reduce inefficiencies in data center operations, support sustainable computing, and maintain service reliability at scale. Experimental results demonstrate that the proposed approach outperforms conventional RL-based models by significantly reducing energy usage and minimizing service-level agreement (SLA) violations. These findings highlight the potential of AI-driven, hybrid scheduling mechanisms for efficient cloud resource management.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Optimized reinforcement learning consolidation agent (ORLCA) framework for intelligent cloud resource management using hybrid MBFD K-Means-CS approach

  • 期刊论文

  • 发表日期：

    2025年11月21日

查看全文
  • Puja Thakur,
  • Arvind Kumar,
  • Jagpreet Sidhu

摘要

Energy efficiency remains a significant challenge in cloud computing, particularly due to the geographically distributed nature of cloud resources. Virtual Machine (VM) consolidation has emerged as a key strategy for optimizing resource utilization and reducing energy consumption. This paper presents a hybrid approach that combines Reinforcement Learning (RL), swarm-based metaheuristics, and the Modified Best Fit Decreasing (MBFD) scheduling algorithm to address energy-aware VM allocation and consolidation in cloud data centers. The proposed method integrates clustering techniques such as K-means and Cosine Similarity (CS) to enhance workload classification and decision-making. By focusing on software-based solutions and leveraging adaptive learning agents, the proposed method aims to reduce inefficiencies in data center operations, support sustainable computing, and maintain service reliability at scale. Experimental results demonstrate that the proposed approach outperforms conventional RL-based models by significantly reducing energy usage and minimizing service-level agreement (SLA) violations. These findings highlight the potential of AI-driven, hybrid scheduling mechanisms for efficient cloud resource management.