Deep reinforcement-based cloud resource allocation based on variable auto-encoder (VAE)

Abstract

The cloud gives minimal effort and adaptable IT resources (equipment and programming) over the Internet. Due to the availability of cloud vendors look to drive more prominent business results and the situations of the cloud become increasingly confounded, through which we can sense that the era of the smart cloud has arrived. The smart cloud faces a few difficulties, including upgrading the monetary cloud administration arrangement and adaptively allotting resources. Specifically, there is a developing pattern toward utilizing AI to improve the knowledge of cloud the executives. This article talks about a design of astute cloud resource the executives with deep reinforcement learning based on auto-encoder. The deep reinforcement learning makes clouds naturally and proficiently arranges the most suitable design, legitimately from entangled cloud situations. At long last, we give a guide to assess and close the amazing capacity of the smart cloud with deep reinforcement learning. We used CloudSim for implementation as a result to increase the effectiveness of proposed method.