Using Distributed Reinforcement Learning for Resource Orchestration in a Network Slicing Scenario

The Network Slicing (NS) paradigm enables the partition of physical and virtual resources among multiple logical networks, possibly managed by different tenants. In such a scenario, network need to be dynamically allocated according slice requirements. this paper, we attack above problem exploiting Deep Reinforcement Learning approach. Our framework is based on distributed architecture, where agents cooperate towards common goal. agent training carried out following Advantage Actor Critic algorithm, which permits handle continuous action spaces. By means extensive simulations, show that our approach yields better performance than both static allocation system an efficient empirical strategy. At same time, proposed ensures high adaptability scenarios without for additional training.