Otimização de processos de alocação e handover em redes móveis

Abstract

The growing number of devices connected to the Internet has required advances in wireless communication technologies. As a result, 5G networks gradually replace 4G and its predecessors, offering more speed, heterogeneity, and scalability. The fifth-generation provides broad support for software-defined networking (SDN) applications, increasing the programming flexibility of processes and protocols previously embedded and difficult to update. This work aims to improve processes in mobile networks through mathematical models. Our work focuses on optimizing the allocation of users in base stations of telecommunication networks, minimizing the handover of users between base stations, and improving network communication quality. The contributions of this work are: i) A mathematical model for allocating users of mobile networks at base stations, also aiming handover reduction; ii) A metaheuristic solution as an alternative to exact models since exact models can prove to be non-scalable and present unfeasible solving times under computationally restricted conditions; iii) A model evaluation in simulated mobility scenarios considering the handover process and the network user distribution according to available bandwidth. Our allocation model considers the average handover frequency of each base station and the Reference Signal Received Quality (RSRQ) indicator between users and base stations. The model evaluation used exact and heuristic methods: the branch and bound algorithm, iterated local search, and a greedy solution. On average, the iterated local search algorithm obtained an execution time reduction of approximately 82% compared to the branch and bound exact algorithm. Regarding the RSRQ indicator, the solution reached a 1.45% average gain, and the number of performed handovers was maintained, compared to a similar literature model. Despite the modest improvement, which makes our proposal statistically equivalent to the literature model, we offer the advantage of not predicting the users’ possible and future routes. Only the current position is required. Furthermore, our solution also considers base stations’ bandwidth capacity, controlling the allocation and network occupation limits.