Massive MIMO joint beamforming and power allocation via LMS with narwhal lemming optimization and fractional ResNeXt-based control

Nowadays, next-generation wireless communication networks are used to provide robust and accurate sensing capability, high-quality wireless connections using massive devices, and support high data transmission rates. Beamforming is used in Massive Multi-Input and Multi-Output (mMIMO) systems to effectively cancel the complex interference caused by the non-zero inner product among the transmitted and channel signal matrices. Various techniques are used to significantly maximize the energy efficiency and reduce the Signal-to-Interference-plus-Noise Ratio (SINR) communication requirements. Meanwhile, these techniques did not maximize the sum-rate of all users by jointly optimizing the transmit beamformer. In this paper, a ResNeXt with Fractional Narwhal Lemming Optimization (ResFraNLO) is introduced for joint beamforming and power allocation in mMIMO systems. Here, the adaptive beamforming is performed using Least-Mean-Square (LMS) adaptive algorithm and the weights of LMS is adjusted optimally using Narwhal Lemming Optimization (NLO). After that, the power allocation is executed using the ResNeXt framework. Moreover, the performance of ResNeXt is improved by fine-tuning the weights and bias by Fractional Narwhal Lemming Optimization (FraNLO). Further, the supremacy of ResFraNLO is investigated, and the ResFraNLO attained achievable sum rate of 79.099 Mbps/Hz, Root Mean Squared Error (RMSE) of 1.34 × 10^-7, and Bit Error Rate (BER) of 5.71 × 10^-9.