A new multi-subpopulation co-evolutionary genetic algorithm for cloud resource scheduling

Abstract

Cloud resource scheduling has become a hot spot in the field of cloud computing. The common method to deal with resource scheduling problem is to build an optimization model and then use intelligent optimization algorithm to solve it. However, only tenant task execution time and the consumed resources in the built models are considered, the improvement of cloud resource scheduling performance is limited. To this end, this paper proposes a new Multi-subpopulation Co-evolutionary Genetic Algorithm called MC-GA for cloud resource scheduling, which comprehensively considers tenant task execution time, task transmission time and resources consumed by task execution. The MC-GA algorithm selects individuals with large fitness values into the selected sub-population through the elitism strategy, and randomly shrinks the size of the sub-population. Experimental results show that the proposed algorithm has the characteristics of higher solution efficiency and stronger optimization ability than other genetic algorithms.