Comparative Study on Cloud Resource Scheduling Using Chaotic Tuna Swarm Optimization Algorithm

As the current standard platform for cloud resource management, Kubernetes generally adopts metaheuristic algorithms for pod scheduling by many researchers due to various shortcomings of its default mechanism. Tuna swarm optimization (TSO) is used as the basic algorithm herein, and a chaotic mapping based population initialization scheme is introduced to address the common problems of metaheuristic algorithms, such as susceptibility to initial populations and premature convergence during later iterations. A comparative study on cloud resource scheduling based on the chaotic tuna swarm optimization (CTSO) using various chaotic maps combined with other improvement strategies, such as levy flight and t-distribution, is conducted in this paper. A series of comparative experiments elaborate the introduced population initialization scheme based on chaotic maps and other improved strategies can effectively improve the optimization performance of the original TSO algorithm.