Enhancing Sum Spectral Efficiency and Fairness in NOMA Systems: A Comparative Study of Metaheuristic Algorithms for Power Allocation

Wireless networks are indispensable in modern life. Nevertheless, these networks are finding it difficult to meet user’s quality of service (QoS) demands due to the sharp rise in the number of network connected devices. Researchers worldwide are investigating novel methodologies to address network demands. Statistics predict that, by 2030, at least 125 billion wireless network-connected devices will generate 5016 EB/month of data traffic with a user density of 107 devices/km2 [1]. The spectral efficiency (SE) requirement for future networks is forecasted to be 100 bps/Hz. The energy efficiency (EE) is targeted to be 100 times higher than that of fifth-generation (5G) with less than 0.1 ms end-to-end delay [2]. Despite their widespread use, orthogonal multiple access (OMA) techniques have limitations. These techniques are designed to enhance the system’s performance but face SE and user data rate restrictions. In OMA, resources, such as time, frequency, or code, are orthogonally allocated among all users. However, its rigid structure prevents it from adapting to the changes in channel conditions, and with more users, OMA cannot fulfill resource demands. As a result, they cannot meet the massive connectivity demand of next-generation networks. This scenario significantly reduces the sum-SE (SSE) of wireless systems [3]. The challenge is to maximize the SSE while simultaneously ensuring fairness and QoS for individual users.

In our work, the utilization of non-orthogonal multiple access (NOMA) is emphasized, and it is widely recognized for providing high SSE in radio access systems. The third generation partnership project (3GPP) has incorporated NOMA, also known as multiuser superposition transmission (MUST), into the long-term evolution advanced (LTE-A) standard [4]. The NOMA-aided downlink provides a new dimension by differentiating the users in the power domain according to their position in the network. NOMA improves SSE by enabling multiple users to utilize the same spectrum concurrently. It uses techniques like superposition coding (SC) at the transmitter and successive interference cancellation (SIC) at the receiver to make more efficient use of the available spectrum compared to OMA. For example, in an OMA system, two users might be assigned different frequency bands, each with a specific bandwidth. In contrast, NOMA allows these two users to use the same band, effectively doubling the potential SSE. NOMA ensures fairness among users by dynamically adjusting power allocation (PA) factors based on channel conditions. Users with weaker channels receive higher power to maintain an acceptable QoS, while stronger users receive lower power but still perform well due to their better channel conditions. NOMA also supports massive connectivity by enabling multiple devices to access shared resources. This is particularly useful for scenarios like the Internet of Things (IoT), where many devices need to communicate at the same time. In an IoT network with numerous sensors, NOMA enables these sensors to transmit data simultaneously over the same frequency band, significantly increasing the number of devices that can be supported compared to an OMA scheme where each device would require a unique resource allocation. Through SC each user will receive the composite signal, and except for the farthest user from the base station (BS), all other users will perform SIC to decode their own data. Hence, NOMA reduces user access time to radio resources, resulting in minimum latency, essential for real-time applications. The features of NOMA significantly enhance its reliability as a solution for massive connectivity.

Effective power distribution in NOMA systems through PA factor optimization improves SSE. The optimization framework generally considers the constraints of maximum transmit power, the user’s SE demands, and channel conditions. The performance of the NOMA system is subject to dynamic channel conditions that can affect the SE of each user. The channel gain may introduce fading in the signal, leading to attenuation and lower data rates. This can be overcome by assigning each user a specific power value that corresponds to their instantaneous channel gains through optimizing PA factors. Considering a large set of instantaneous gains and corresponding PA factors will enhance the system’s statistical average performance with maximum SSE. It is important to consider the probability of outage of user devices when optimizing the PA factors.