Resource Scheduling for Offline Cloud Computing Using Deep Reinforcement Learning

Many organizations around the world use cloud computing for their services. Cloud Computing is mainly based on the concept of on-demand delivery of computations, storage, applications, and other resources. It depends on delivering users services through Internet connectivity. It also uses a pay-as-you-go business model to handle users' services. It has some essential characteristics including on-demand service, resource pooling, rapid elasticity, virtualization, and measured services. At the same time, there are different types of virtualization such as full virtualization, paravirtualization, emulation, OS virtualization, and application virtualization. Resource scheduling in the cloud computing is one of the most challenging jobs where resources have to be allocated to the required tasks/jobs according to the required Quality of Services (QoS) of the cloud applications. Due to the cloud environment, uncertainty, and maybe heterogeneity, resource allocation cannot be addressed with the existing policies. The problem still a major concern of most of the cloud providers where they face troubles in selecting the appropriate resource scheduling algorithm for a specific workload, especially the workload might be dynamic. In this paper, we use one of the Artificial Intelligence (AI) emergent algorithms, Deep Reinforcement Learning, (Deep Reinforcement Learning for Cloud Scheduling (DRLCS)), to solve the problem of resource scheduling in cloud computing.