Power Allocation in Cache-Aided NOMA Systems: Optimization and Deep\n Reinforcement Learning Approaches

This work exploits the advantages of two prominent techniques in future\ncommunication networks, namely caching and non-orthogonal multiple access\n(NOMA). Particularly, a system with Rayleigh fading channels and cache-enabled\nusers is analyzed. It is shown that the caching-NOMA combination provides a new\nopportunity of cache hit which enhances the cache utility as well as the\neffectiveness of NOMA. Importantly, this comes without requiring users'\ncollaboration, and thus, avoids many complicated issues such as users' privacy\nand security, selfishness, etc. In order to optimize users' quality of service\nand, concurrently, ensure the fairness among users, the probability that all\nusers can decode the desired signals is maximized. In NOMA, a combination of\nmultiple messages are sent to users, and the defined objective is approached by\nfinding an appropriate power allocation for message signals. To address the\npower allocation problem, two novel methods are proposed. The first one is a\ndivide-and-conquer-based method for which closed-form expressions for the\noptimal resource allocation policy are derived, making this method simple and\nflexible to the system context. The second one is based on the deep\nreinforcement learning method that allows all users to share the full\nbandwidth. Finally, simulation results are provided to demonstrate the\neffectiveness of the proposed methods and to compare their performance.\n