On optimizing resource allocation

Resource allocation remains a challenging issue in communication networks, and its complexity is continuously increasing with the densification of the networks. With the evolution of new wireless technologies such as Fifth Generation (5G) and Sixth Generation (6G) mobile networks, the service level requirements have become stricter and more heterogeneous depending on the use case. In this paper, we review a large body of literature on various resource allocation schemes that are used in particular in mobile wireless communication networks and compare the proposed schemes in terms of performance indicators as well as techniques used. Our review shows that among the strategies proposed in the literature, there is a wide variety of optimization targets and combinations thereof, focusing mainly on performance indicators such as energy efficiency, spectral efficiency, and network capacity. In addition, in this paper, selected algorithms for resource allocation are numerically analyzed through simulations to compare and highlight the importance of how the resource algorithms are implemented to achieve efficient usage of the available spectrum. The performance of selected algorithms is evaluated in a multi-cell heterogeneous network and compared to proportional fair and eICIC, a widely-used combination of resource allocation and interference mitigation techniques used by communication networks. The results show that one approach may perform better when looking at the individual average user data rate but worse when looking at the overall spectral or energy efficiency, depending on the category of traffic. The results, therefore, confirm that there may not be a single algorithm that visibly outperforms other candidates in terms of all performance criteria. Instead, their efficiency is always a consequence of a strategic choice of goals, and the targeted parameters are optimized at a price. Thus, the development and implementation of resource allocation algorithms must follow concrete usage scenarios and network needs and be highly dependent on the requirements and criteria of network performance.