QoS Multicast Routing Optimization Research Based On Genetic Algorithm

With the advent and the rapid development of the Internet, there are more and more multimedia applications constantly emerging, such as vedio conferencing, remote video teaching and vedio on demand. Although all these applications have different forms, they all have one thing in common: for a higher demand to efficient transmission of information. In order to better ensure the transmission quality of these services, there is a certain requirement for the information to meet, one that is called the Qo S(Quality of Service) issues. Multiple constrained Qo S problem has been proved to be an NP-hard problem, while the traditional routing protocols provide a “best effort” service which can not effectively guarantee Qo S. In order to effectively meet the Qo S constraints, intellegent algorithms are used to solve multiple Qo S constraints NP-hard problem, and in many intellegent alogrithms, the genetic alogrithm is featured by better global search parallelism and self adaptability,which has more advantage to solve NP problems. Now there are many studies about multiple Qo S multicast routing based on genetic algotithm, but two major issues must be solved on how to effectively combine multicast routing problem and genetic algorithm: one is how to set the population of individuals to complete coding, the other is how to set a good fitness function. Good coding can effectively complete the crossover and mutation without decoding. Genetic algorithm just needs to be assessed according to the individual fitness function value,without any knowledge and extra information of the search space. By genetic operations we can get the calculated solution.To solve the above problems, this paper gradually researches multiple Qo S unicast routing, the multicast routing which meets delay and bandwidth constraints and multiple Qo S multicast routing. Finally three improved genetic algorithms have been approached for each routing problem.The first algorithm is mainly set up a fitness function with a new punishment mechanism, the second algorithm major made an innovative encoding, and the last one is an integrated algorithm for the first two algorithms. Simulation show that the proposed algorithm is feasible, and by comparing the existing genetic algorithms and the proposed algorithms, the result has been proved article alogrithms with better performance such as better convergence、less resource consumption and smaller cost.