A Cloud Computing Resource Scheduling Policy Based on Genetic Algorithm with Multiple Fitness

Under the cloud computing environment of IaaS(Infrastructure as a Service), due to the expansion of system scale and virtual machines' migrations, etc, it is easy to cause some problems like fragmentation of physical resources, low utilization of resources. The consequences lead to high energy consumption within an Internet Data center. In this paper, we propose a pre-migration strategy based on three load dimensions: CPU utilization, network throughput, disk I/O rate, which are considered complementary in the algorithm. In order to get an approximately optimal solution, we adopt the hybrid genetic algorithm combined with knapsack problem with multiple fitness and experiments are conducted to verify the effectiveness of the algorithm. The result of the experiments shows that the algorithm can effectively achieve the goal of raising resources' utilization and lowering energy consumption under cloud computing environment.