Optimisation des paramètres d’un réseau Ad Hoc par algorithmes génétiques

Abstract

The problem of routing in Ad Hoc networks, in particular proactive routing, hasattracted the attention of . Although the protocols proposed in the literature present some relevant characteristics, they have limitations, especially highly mobilenodes or high network load-dependent parameters of the Ad Hoc network often chosen intuitively by a seasoned expert. This thesis describes our work to solve this difficult task using a genetic algorithm for automatic parameter selection. The use of multi-objective genetic algorithm can not only minimize the average delay of transferof the network as well as the number of lost packets but also to adapt to each change of topology which makes it adaptive to any environment.