A Comparative Study of Optimization Based Task Scheduling in Cloud Computing Environments Using Machine Learning

Cloud computing has brought about a significant transformation in the field of Information Technology (IT) by providing the capability to access a shared pool of resources over the internet in a scalable and on-demand manner. For this purpose, task scheduling and balancing play an impactful role in the cloud since the efficient scheduling of tasks encounters issues arising from the variability of workloads, the availability of resources, and to optimize both performance and cost. This study presents a comparative study of optimization-based task scheduling mechanisms in cloud computing with the help of supervised Machine Learning method. Here, we performed several experiments on several swarm-based optimization approaches to solve the scheduling process in a balanced manner with the help of Artificial Neural Network (ANN) as a supervised machine learning mechanism. Firstly, we perform the task scheduling using the traditional approach, and after that, we apply optimization mechanism to improve the efficiency of the model based on the problems that occur during the allocation of resources and balancing mechanisms. So, to analyze the complete scheduling mechanism and problem identification mechanism, the concept of Back-Propagation Algorithm (BPA) with ANN is used that is named as ANN-BPA. For the selection of optimal features to train the ANN -BP A, the concept of swarm-based optimization method is used for random simulation -based data. At the end, to validate the model efficiency, a comparative analysis is performed with several swarm algorithms like Particle Swarm Optimization (PSO), Artificial Bee Colony (ABC), Cuckoo Search Optimization (CSA) and Moth Flame Optimization (MFO). Here, the combination of MFO with ANN-BPA performs better than the other in terms of task allocation, completion, execution time, andenergy consumption.