In this work, the primary objective is the optimization of routing in ad hoc networks big map, which we studied through two different aspects of artificial intelligence. In the first part, we propose the Genetics algorithm (AG) as a solution to resolve the np-difficult problem of routing optimization in ad hoc networks big map. In the second part, we propose Ant Colony algorithms (ACO) to solve the unicast routing problem for quality of service (QoS) improvement by using parallel programming to significantly reduce the computation time required by this type of problem. The research studies studied in the framework of this thesis, can be seen as an answer to the following question: What is the best approaches can be adopted to optimize large ad hoc networks? One of the constraints in sensor networks: is the problem transmitting data from a source to an estimate, passing through existing sensors in the network with a real need to guarantee a better quality of service (QoS) during the rapid, evolutionary and unpredictable change of the topology of this network whose nodes are in constant motion. Our problem, called unicast routing problem with QoS in the literature, is an NP-difficult problem that cannot be solved by traditional exact methods. For this reason, several meta-heuristic algorithms have been used to solve it in order to satisfy the constraints of delay, bandwidth and packet loss rate. In this work, we propose two new ideas based on artificial intelligence integrating meta-heuristic methods (AG) et (ACO), machine learning and parallel programming to achieve good results defying those of the literature in terms of calculation time and quality of the solution. Thus, the comparison of these two methods will help the user to choose later the best particular method, adapted to the networks of ad hoc vehicles studied.