Principal Component Regression-Based Adaptive Multiple Extrema Seeking Cat Swarm Resource Optimized Task Scheduling In Cloud Computing

Abstract

In this paper, task scheduling process is a challenging task in cloud computing to determine the best optimal virtual machine for each task. Many types of scheduling algorithms have been introduced for small or medium-scale cloud computing. However, dynamic scheduling is a major challenging problem for large-scale cloud computing environments. To address the issue, this paper proposes a novel technique called Principal Component Regression-based Adaptive Multiple Extrema Seeking Cat Swarm Resource Optimization (PCR-AMESCSRO) technique for efficient task scheduling with lesser makespan and higher efficiency. The PCR-AMESCSRO technique is designed with the contribution of Principal Component Regression (PCR) and Adaptive Multiple Extrema Seeking Cat Swarm Optimization Algorithm (AMESCSOA). First, the PCR is applied to analyze the user requested task and assign the priority level with lesser makespan. Second, the AMESCSOA is used to identify the optimal virtual machine by the cloud manager. Lastly, the experimental valuation is performed on factors such as task scheduling efficiency, false-positive rate, makespan, and memory consumption with respect to a number of user tasks. The observed results show the superior performance of our proposed PCR-AMESCSRO technique when compared to state-of-the-art methods.