GR-314 Reinforcement Learning based Offloading Scheme Computation to Optimize Latency-Energy in Collaborative Cloud Networks

GR-314 Reinforcement Learning based Offloading Scheme Computation to Optimize Latency-Energy in Collaborative Cloud Networks

Growing technologies like virtualization and artificial intelligence have become more popular on mobile devices. But lack of resources faced for processing these applications is still a major hurdle. Collaborative edge and cloud computing are one of the solutions to this problem. Remote servers have enough resources to support computation-heavy tasks and compute the results faster. But transmission time and energy are involved while offloading the computation to remote servers such as cloud and edge devices. There is a need to find an optimal offloading ratio for cloud as well as edge servers such that entire computation on remote as well as local can be achieved minimum energy consumption as well as minimum delay. We have proposed a multi-period deep deterministic policy gradient (MP-DDPG) algorithm to find an optimal offloading policy by partitioning the task and offloading it to the collaborative cloud and edge network to reduce energy consumption. Our results show that MP-DDPG achieves the minimum latency and energy consumption in the collaborative cloud network. We have compared our results with the existing DDPG-based approach and achieved about 65% speedup in terms of latency. Also, we observed energy consumption reduces with an increase in the number of edge servers.