Scholarlink

Cloud offers a fantastic environment for users to process their requests by submitting tasks for computations. This platform produces supreme Quality of Service (QoS) by processing tasks using Resource Scheduling Algorithms (RSA) on its Virtual Machines (VMs). The cloud can give ideal results by adequately managing the VMs managed by these RSAs. This experimental research mainly aims to analyze the behavior of these cloud RSAs using mathematical regression models. In this experimental research, tasks of various smaller and larger sizes are computed across four experimental scenarios using RSAs, namely First Come, First Serve (F_C_F_S), Maximum—Minimum (Mx_Mn), Minimum Completion Time (M_C_T), Minimum—Minimum (Mn_Mn), and Round Robin (R_R). The cost obtained to compute these tasks is recorded in each experimental scenario and later provided as an input to the two mathematical regression models, namely Exponential and Logarithmic to compare their performance mathematically. The experimental results convey that for smaller task sizes, the Mx_Mn provides ideal results, and R_R provides poor results. Further, for comparatively large task sizes, the Mn_Mn provides ideal results, and M_C_T and R_R provide poor results. Lastly, an intelligence mechanism is suggested using Reinforcement Learning to hybridize the cloud RSAs and improve the performance of the cloud environment.

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

Exponential and Logarithmic Regression Models to Improve Cloud Performance Using Reinforcement Learning

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Prathamesh Vijay Lahande,
  • Parag Ravikant Kaveri,
  • Shirish Chintaman Joshi

摘要

Cloud offers a fantastic environment for users to process their requests by submitting tasks for computations. This platform produces supreme Quality of Service (QoS) by processing tasks using Resource Scheduling Algorithms (RSA) on its Virtual Machines (VMs). The cloud can give ideal results by adequately managing the VMs managed by these RSAs. This experimental research mainly aims to analyze the behavior of these cloud RSAs using mathematical regression models. In this experimental research, tasks of various smaller and larger sizes are computed across four experimental scenarios using RSAs, namely First Come, First Serve (F_C_F_S), Maximum—Minimum (Mx_Mn), Minimum Completion Time (M_C_T), Minimum—Minimum (Mn_Mn), and Round Robin (R_R). The cost obtained to compute these tasks is recorded in each experimental scenario and later provided as an input to the two mathematical regression models, namely Exponential and Logarithmic to compare their performance mathematically. The experimental results convey that for smaller task sizes, the Mx_Mn provides ideal results, and R_R provides poor results. Further, for comparatively large task sizes, the Mn_Mn provides ideal results, and M_C_T and R_R provide poor results. Lastly, an intelligence mechanism is suggested using Reinforcement Learning to hybridize the cloud RSAs and improve the performance of the cloud environment.