… a deep reinforcement learning (DRL) agent that … -numerology interference (INI). Furthermore,
by exploiting the information about the deterministic traffic patterns, we specialize the agent …

… In addition, we boost the agent learning efficiency by designing an action masking module
… for a mixed-numerology access scheme. The authors of [8] design a DRL agent to allocate …

… agent-based allocation of mixednumerology spectrum slices by leveraging deep reinforcement
learning … slices that are multiplexed on a mixed-numerology physical layer. The objective …

… for a multi-user scenario in a mixed-numerology 5G RAN shared by different network slices.
… 6, we show the agent capability to gradually learn the feasible action space. Specifically, we …

… Q-learning focuses on how agents interact with the unknown environment to maximize Q …
So, we will combine Qlearning with deep learning called Deep Q-Learning to find the optimal …

… The NR frame structure introduced mixed numerology where each numerology represents …
Action (A): based on the current state st∈ S, the learning agent selects the best action at ∈ A …

… Learning (DRL) to allocate resources and numerology for UEs to … as a consequence of
mixed numerologies. However, the … Then, we will evaluate the trained agent in a 5G simulated …

… , mixed-numerology access schemes on the shared RAN physical layer further exacerbate
the policy design complexity since Inter-Numerology … Furthermore, the designed agents are …

… , in this case market learning appears to be on the sell-side of the market: inexperienced
agents state WTA values more than three times larger than experienced agents, whereas WTP …

… learning (DRL) is a powerful machine learning tool to deal with discrete decision-making
problems, we provide a DRL-based solution, in which an agent … on mixed numerology and mini-…