Treffer: A Comparative Study of Cloud Job Scheduling Algorithms with a Hybrid Genetic and Biogeography-based Optimization Algorithm.

A hybrid cloud is a platform that enables the owner of any business house to procure specific local resources and various other resources from some external service provider. This model is quite helpful in avoiding as well as handling the condition of a "cloud burst", where a private cloud setup becomes overburdened with processing load and then immediately has to switch to a public cloud to deal with the situation. Cloud Scheduling tries to achieve the "most eminent" plan or schedule for data centers to handle demanded services with minimum communication losses while incurring the least production cost. The objective of this research paper is to compare the hybrid Improvised Biogeography-Based Optimization-Genetic Algorithm, developed to enhance the efficiency of the process of cloud job scheduling. The method consisted of incorporating the functions of mutation and crossover from a genetic algorithm into the functions of the Biogeography-Based Optimization algorithm. The scheduling cost and throughput of the Improvised-Biogeography-Based Optimization-Genetic Algorithm were compared on an open cloud raw dataset using MATLAB with five existing algorithms, namely, Genetic Algorithm, Ant Colony Optimization, Biogeography-Based Optimization, Grey Wolf Optimization and Whale Optimization algorithms. The Improvised-BBO-GA was found to be more efficient as compared to all the other five algorithms in terms of a significantly reduced scheduling cost and a much higher throughput. Based on multiple runs of all the algorithms, the width of the confidence interval was found to be the narrowest for the proposed algorithm. [ABSTRACT FROM AUTHOR]