Planification globale des reseaux mobiles de la quatrieme generation (4G)

In the current context where information is the key to success in any field where one stands, telecommunications networks are increasingly in demand. Huge amounts of information circulates on the networks every second. It is essential to ensure the availability of these networks to ensure the transmission of these data at any time.;The problem of planning of telecommunication networks is to determine, from a set of potential sites, those to be used to cover a given geographical area. One should also choose the equipment to be installed on these sites and to link them according to certain well-defined constraints. For decades, several authors have focused on solving this problem in order to minimize the cost of network installation. These authors were interested in various aspects of the problem without considering it in its entirety.;Some studies have recently been performed on the global planning of mobile networks. The authors were interested in the third generation networks. They proposed a model to solve the problem entirely, without breaking it down into sub-problems. However, they did not take into account the fault tolerance of network.;This thesis proposes a global planning framework for the fourth generation (4G) networks (the new generation of mobile networks). The survivability of the network is taken into account in this study. The work was conducted in three phases.;In the first phase, a global model including survivability has been designed for the planning of 4G (WiMAX) networks and solved optimally with a mathematical solver using the integer linear programing method. The objective of the model is to minimize the network cost while maximizing its survivability. To show the relevance of the global resolution, the model was compared to a sequential model with the same constraints. The sequential model is to divide the problem into three sub-problems and solve them successively. A global model which does not include survivability constraints has also been designed to test the effect of failures on the network. The results show that the proposed model performs on average 25% better than the two other models.;The problem of global network planning and the problem of survivability of telecommunications networks are two NP-hard problems. The combination of these two problems provides a problem even more difficult to solve than each problem taken separately. The exact method used in the first phase can only solve small instances. In the second phase, we propose a hybrid metaheuristic to find 'good solutions' in a 'reasonable time' for instances of larger size. The proposed metaheuristic is a new form of hybridization between the iterated local search algorithm and the integer linear programing method. The hybridization of these two methods can benefit from their respective advantages, namely the efficient exploration of the search space and the intensification of the solutions obtained. The intensification is performed by the exact method that calculates the best possible solution from a given configuration while the exploration of the search space is made through the iterated local search algorithm. The performance of the algorithm have been evaluated with respect to the exact method given in the first phase. The results show that the proposed algorithm generates solutions that are on average 0,06% of the optimal solutions. For the larger instances, the lower bounds are calculated using a relaxation of the model. The comparison of the results obtained by the proposed algorithm with the lower bounds show that the metaheuristic obtains solutions that are on average 2,43% from the lower bounds, for the instances that cannot be solved optimally, within a much less computation time.