A Deep Learning With Optimization-Based Power Allocation for Network Slicing in MIMO–NOMA

Abstract

Nonorthogonal multiple access (NOMA) and multiple-input multiple-output (MIMO) are regarded as the best technologies for handling high-rate requirements. Nevertheless, the consumption of energy for huge amounts of chains leads to issues in energy efficiency (EE) requirements. Therefore, a new technique has been introduced for enhancing power allocation (PA). Initially, the system model for network slicing (NS) is considered, and then, quadrature amplitude modulation (QAM) is executed for transmitting the information. Next, orthogonal frequency division multiplexing (OFDM) is performed to divide the radio channel into many closely spaced subchannels. Then, preamble insertion is executed for channel equalization and data synchronization, and NS with massive MIMO is implemented for allocating communication resources to users. At last, PA is done by back propagation neural network (BPNN) by considering various parameters. Here, BPNN is tuned by harmonic ladybug beetle honey badger optimization (HLBHBO), where HLBHBO is formulated by combining harmonic analysis and ladybug beetle honey badger optimization (LBHBO). Moreover, LBHBO is engineered by the amalgamation of the honey badger algorithm (HBA) and ladybug beetle optimization (LBO). The experimental outcomes of HLBHBO + BPNN attained the highest sum rate of 1.990 Mbits/s, EE of 19.572 bits/J, and achievable rate of 149.857 Mbits/s.