Scholarlink

<p>Cloudlet placement in dynamic environments poses significant challenges due to the mobility of users. Effective placement strategies are crucial for enhancing the performance of mobile applications. Recent research has applied clustering algorithms to tackle these challenges, including <i>k</i>-means, mini-batch <i>k</i>-means, and <i>k</i>-medoids. This paper introduces an enhanced clustering and filtering cloudlet movement and placement (<i>e</i>-CFCMP) algorithm designed to identify optimal positions for movable cloudlets in dynamic environments, thereby supporting the efficient operation of mobile applications. The proposed algorithm includes three variants: <i>e</i>-CF<i>km</i>, <i>e</i>-CF<i>mb</i>, and <i>e</i>-CF<i>kd</i>. Among these, the first variant achieves the optimal position by maximizing the number of covered devices while minimizing assignment costs. It outperforms the other two variants, covering 0.88% and 1.51% more devices, respectively. Additionally, it reduces intra-cluster assignment costs by 1.58% and nearly 99% compared to contemporary algorithms. Furthermore, the algorithm ensures 100% placement of cloudlets with uniformly distributed devices, enhancing resource utilization. Therefore, the <i>e</i>-CF<i>km</i> variant performs better than its counterparts.</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

Enhanced clustering and filtering algorithm for cloudlet movement and placement in dynamic mobile cloud computing environments

  • 期刊论文

  • 发表日期：

    2025年11月13日

查看全文
  • Lizia Sahkhar,
  • Bunil Kumar Balabantaray,
  • Sanjaya Kumar Panda,
  • David Taniar

摘要

Cloudlet placement in dynamic environments poses significant challenges due to the mobility of users. Effective placement strategies are crucial for enhancing the performance of mobile applications. Recent research has applied clustering algorithms to tackle these challenges, including k-means, mini-batch k-means, and k-medoids. This paper introduces an enhanced clustering and filtering cloudlet movement and placement (e-CFCMP) algorithm designed to identify optimal positions for movable cloudlets in dynamic environments, thereby supporting the efficient operation of mobile applications. The proposed algorithm includes three variants: e-CFkm, e-CFmb, and e-CFkd. Among these, the first variant achieves the optimal position by maximizing the number of covered devices while minimizing assignment costs. It outperforms the other two variants, covering 0.88% and 1.51% more devices, respectively. Additionally, it reduces intra-cluster assignment costs by 1.58% and nearly 99% compared to contemporary algorithms. Furthermore, the algorithm ensures 100% placement of cloudlets with uniformly distributed devices, enhancing resource utilization. Therefore, the e-CFkm variant performs better than its counterparts.