Deep Reinforcement Learning based Resource Optimization in HCRAN

Heterogeneous Cloud Radio Access Networks (HCRANs) are now the prime paradigm in fulfilling the ever-growing requirements of wireless communications, especially for the 5G and beyond era. Resource allocation, namely spectrum and power allocation, is at the center of realizing optimum network performance and user experience. It necessitates careful decision-making to serve diverse needs of many users and services with a minimum of disturbance and a maximum of throughput. Traditional resource allocation algorithms, as helpful as they are in static and predictable systems, risk losing control over the dynamic and complex HCRAN environment. Deep reinforcement learning (DRL) appears to be an overwhelming prospect. DRL utilizes the power of artificial intelligence for adaptive and intelligent resource allocation decision in a dynamic HCRAN scenario. In this article we explore potential DRL with multiagent support based algorithm to optimize resource allocation in HCRAN networks.