Dividing the transmission power statically among Long Term Evolution (LTE) and New Radio (NR) carriers during dual connectivity may result in a great deal of unused power, which may lower the throughput and user experience. RAN4 has mandated the need for Dynamic Power Sharing (DPS) for all the User Equipments (UEs) supporting LTE-NR dual connectivity. DPS allows the UE to distribute the transmit power among carriers. However, optimizing the power distribution among the LTE and NR carriers during dual connectivity efficiently and effectively is still an open problem. In this paper, we propose a novel distributed Q-learning assisted DPS scheme to learn the distribution of transmission powers for LTE and NR carriers for dual connectivity. Extensive simulations and analysis reveal that our proposed scheme improves the network performance by allocating transmission powers optimistically considering channel fluctuations.

https://ieeexplore.ieee.org/document/9045097