A Stochastic Approximation Approach for Foresighted Task Scheduling in\n Cloud Computing

With the increasing and elastic demand for cloud resources, finding an\noptimal task scheduling mechanism become a challenge for cloud service\nproviders. Due to the time-varying nature of resource demands in length and\nprocessing over time and dynamics and heterogeneity of cloud resources,\nexisting myopic task scheduling solutions intended to maximize the performance\nof task scheduling are inefficient and sacrifice the long-time system\nperformance in terms of resource utilization and response time. In this paper,\nwe propose an optimal solution for performing foresighted task scheduling in a\ncloud environment. Since a-priori knowledge from the dynamics in queue length\nof virtual machines is not known in run time, an online reinforcement learning\napproach is proposed for foresighted task allocation. The evaluation results\nshow that our method not only reduce the response time and makespan of\nsubmitted tasks, but also increase the resource efficiency. So in this thesis a\nscheduling method based on reinforcement learning is proposed. Adopting with\nenvironment conditions and responding to unsteady requests, reinforcement\nlearning can cause a long-term increase in system's performance. The results\nshow that this proposed method can not only reduce the response time and\nmakespan but also increase resource efficiency as a minor goal.\n