부산대학교 도서관

The fifth generation (5G) of mobile networks utilizing millimeter Wave (mmWave) bands can be considered the leading player in meeting the continuously increasing hunger of the end user demands in the near future. However, 5G networks are characterized by high power consumption, which poses a significant challenge to the efficient management of base stations (BSs) and user association. Implementing new power consumption and user association strategies is imperative to address this issue. For that in this work, we focus on the Dual Connectivity-User and Power Allocation (DC-UPA) problem utilizing BS switching on/off along with user dual connectivity. The problem is mathematically formulated as an Integer Linear Program (ILP), and its NP-hardness is proven by showing its equivalence to a variant of the set covering problem. Moreover, we developed two heuristic algorithms: Simulated Annealing (SA) and Distance-Aware (DA) greedy, to mitigate the impact of the problem complexity and resolve the ILP scalability issue. We conducted extensive simulations to validate the effectiveness of the proposed heuristics in a two-dimensional area containing multiple BSs and users with uniform and nonuniform geographical distributions. The performance of the SA and DA algorithms was compared against the ILP approach. We evaluated the performance of the proposed solutions considering different aspects such as the number of users, the BS radius, and the traffic load changes. The numerical results show that SA outperforms the DA in both uniform and nonuniform geographical distributions of users. The SA provides a sub-optimal solution with an optimality gap of about 3.2%, while the optimality gap of the DA is 8.62% in the case of the uniform distribution. Moreover, the optimality gap in the case of nonuniform distribution is equal to about 1% and 5.2% of SA and DA, respectively. Additionally, by utilizing our solutions, the reduction of the level of power consumption up to 16.1% and 20% in the case of uniform and nonuniform distributions can be achieved. The obtained results highlight the efficiency of the proposed algorithms in addressing the DC-UPA problem, providing practical solutions for managing power consumption and maintaining continuous user connectivity in 5G mmWave networks.