Abstract: The shortest path problem is studied for finding the shortest route from a specified source to a specified destination in a mobile wireless networks with minimizing the total cost associated with the path. Several QoS measures are proposed to calculate the best path for routing the packets from specific sources to their destination. These measures are end-to-end delay, bandwidth, and No. of hops. The main goal of this work is to solve the problem of route optimization between the sender and receiver in a dynamic mobile wireless networks, and formulating the multi objective QoS measures in such dynamic environment as a multi- objective optimization using weighted sum approach. A new algorithm based on evolutionary multi objective genetic algorithm technique has been proposed and called Adaptive Genetic Algorithm (AGA) to find out the optimal route in dynamic wireless networks that satisfied the multi objective QoS measures. The proposed algorithm is adaptive in the sense that it finds the shortest route even with the dynamic nature of the mobile wireless networks, e.g. moving nodes and a reproduction operator of the proposed algorithm which uses six different selection methods that are changing through the generations of the proposed algorithm and the best selection method is chosen by AGA according to maximum fitness. An experiment has been made for illustrating the behaviour of our proposed algorithm in wireless networks; in this experiment, the AGA has been implemented on double objectives QoS. The AGA is implemented online with predetermined initial population and fitness values. The simulations have been done under MATLAB and Visual Basic environments, and they showed that our proposed AGA performs excellently and adapts quickly to the dynamic nature of the wireless network and satisfying all of the constraints and objective measures imposed on the networks.
Keywords: Adaptive Genetic algorithm, Quality of Service (QoS), double objective optimization, routing, end-to-end delay, shortest path, dynamic networks.