Association optimale d'utilisateurs dans un réseau cellulaire hétérogène

Abstract

Year by year, the number of cellular network users continues to grow remarkably. This observation is illustrated in several technical reports and statistical studies. On the other hand, users become more and more exigent in terms of resources demands whether in quantity or quality of services. In recent years, it has been found that the application of the conventional network no longer meets the requirements of consumers and is no longer appropriate for the technological developments confronted by today’s world. Thus, industrialists and academicians have developed the idea of exploiting the heterogeneity of cellular networks, where cells with different range are embedded in the same network. The aim of using small cells in conjunction with conventional macrocells is to be able to serve as many users as possible while satisfying the quality of service they require. Moreover, it has been proven that the exploitation of the heterogeneity is less expensive compared to the case where other macrocells are added to the networks. However, HetNets create several complexities and challenges that one must admit. Among them are the cell selection and the load balancing between the macrocell and small cells which is due particularly to the disparity of their transmission power, their coverage and their available resources. A second important challenge is the user mobility management when the user crosses a picocell or femtocell region at a high speed. The third important point to consider is the interference measured by small cells users from macrocells. In this context, the interference is mainly due to the fact that users are pushed to associate with small cells in the presence of neighboring macrocells. Therefore, in this thesis, we try to find appropriate solutions for the three major problems listed above, namely, the cell selection, the intercell handover and the interference management in a HetNet. In the first part, we propose heuristic algorithms to solve the original NP-hard association problem in HetNet. More precisely, two algorithms are studied there. The first algorithm is an association technique based on the dynamic extension of picocell’s coverage. On the other hand, the second user association algorithm relies on the fairness gain measured by the user from the potential neighboring cell. Throughout the analyzes, one observes that, the proposed heuristic algorithms are simple to perform and give a better performance compared to the general methods proposed in the literature in terms of cell overload reduction and equalized distribution of the network’users. In the second part, our goal is to optimize the user fairness in HetNet and to reduce the interference measured by extended femtocells’users. In this perspective, we exploit the technique of cell coverage expansion. Thus, we propose an optimal coordinated cell range expansion technique that compute the optimal cells biases while accounting of several parameters for achieving optimality. We prove that, compared to some of the previously studied methods, the technique we propose results in a better improvement related to the load balancing, the equity between users and the throughput of each user. In the third part, assuming that handover is a technique of user mobility management in cellular networks, we improve the user’s experience during a handover process by considering a scenario where the latter crosses a femtocell coverage. Thus, we propose a utility-based technique whose main idea is the design of new utility functions via a specific objective function that takes account of the user's requirement. Subsequently, the handover bias is deduced and inserted into the transfer decision policies. We conclude that a high gain is obtained in terms of the femtocell assignment probability while maintaining an acceptable number of intercellular handovers.

Date	29 Nov 2017
Original language	French
Awarding Institution	École de technologie supérieure
Supervisor	François Gagnon (Supervisor)