Multi-Faceted Job Scheduling Optimization Using Q-learning With ABC In Cloud Environment

Abstract

Resource allocation is the utmost challenging and common problem, particularly in the cloud service model in Infrastructure as a Service (IaaS). The issue of load balancing is so harmful that irregular load balancing may result in a structure smash. Adopting a suitable access plan and allowing the system to spread work among all existing resources leads to utilizing Virtual Machines (VMs) appropriately. To get enhanced results from the Artificial Bee Colony Algorithm (ABC), a reinforcement learning technique Q-learning is combined using multifaceted job scheduling optimization based on ABC(QMFOABC) has been proposed. The proposed approach improves resource utilization and scheduling created on resource, cost, and makespan. The efficiency of the suggested strategy was evaluated in Datasets Synthetic Workload, Google Cloud Jobs (GoCJ), and Random by using CloudSim to the remaining scheduling strategies for load matching like Max-min, Multifaceted Cuckoo Search (MFCS), Multifaceted Particle Swarm Optimization (MFPSO), Q-learning, Heuristic job scheduling with Artificial Bee Colony approach with Largest Job First algorithm (HABC LJF), First come first serve (FCFS). According to the findings of the experiments, the algorithm that employed the QMFOABC method has better results in resource utilization, throughput, cost, and makespan. Compared to Max-Min (82.31%), MOPSO (35.62%), HABC LJF (21.65%), Q-Learning (11.72%), VTO-QABC FCFS(5.87%), and VTO-ABC LJF (5.86%) shorter time than MOCS, a considerable improvement is found.