High-speed Optimization of User Pairing in NOMA System Using Laser Chaos Based MAB Algorithm

Abstract

NOMA has become one of the most concerned technologies in 5G technology due to its efficiency in improving spectrum utilization by exploiting power domain technology. Efficient pairing techniques are needed to avoid system performance decrease affected by Successive Interference Cancellation (SIC) execution errors in NOMA. Generally, user pairing in NOMA is based on the distance between the Base Station (BS) and the users, which may lead extra time consumption in measuring distance. To reduce the communication latency caused by user pairing, we propose an ultra-fast user pairing algorithm using laser chaos based Multi-Armed-Bandit (MAB) algorithm. It has been shown that ultra-fast and high-performance decision making on the order of GHz can be achieved by applying laser chaos based MAB algorithm. In our proposed algorithm, user pairing is based on all possible pairing combination instead of distance. Simulation results show that our proposed algorithm can achieve higher throughput than the state-of-the-arts NOMA pairing algorithms, i.e., C-NOMA and UCGD-NOMA, with high speed under different radii of communication area.