Sustainable Resource Allocation in Edge environment based on Deep Deterministic-Reinforced Learning

Edge computing has the potential to revolutionise the speed and efficiency with which the information is processed particularly in real-time. Also, Edge settings are dynamic and comprise a wide range of elements, it is found effective handling multiple tasks at any instance. However, in such dynamic situations resource allocation is quite challenging. Resource allocation is expected to be optimistic and tamperproof so that the efficiency and performance of the overall edge computing environment shall not compromised. In this paper, an extensive study about resource allocation is carried and proposed a reinforcement learning based solution to effectively handle the dynamic resource allocation. The model uses a Deep Deterministic Policy Gradient (DDPG) is used to identify the need of resource allocation and apply the allocation strategies. The experiment is conducted using CRAWDAD KAIST - WIBRO dataset, which simulated the real workload and resources data. The obtained results are convincing, especially in terms of resource usage, response time, and energy efficiency.