A Self-adaptive Particle Swarm Mutation Rate Of Genetic Algorithm In Network Coding Link Optimization

Modern society is a complex network system consisting of various information networks.Network optimization and their corresponding specific models are widely used to solve the problems in many fields.Especially,the optimization of network coding links is the specific model in the field of communication,which has a great research value in both theory and economy aspects,including the utilization rate of bandwidth,the improvement of network throughput,the energy consumption of network nodes and network load balance.It is undeniable that network coding has some disadvantages,such as the consumption produced by the coding operation of the intermediate node and increasing the computational complexity of the network,so the optimization of network coding links has fatal research significance.In this thesis,we mainly do some researches about some specific models of network encoding link optimization by adaptive algorithm,the aim is to reduce the cost of encoding while reducing the maximum multicast rate of the network.For the purpose of solving the problem of network encoding link optimization,a mixed optimization algorithm about particle swarm optimization algorithm and the genetic algorithm is put forward in this thesis.In the whole evolution process,we adopt roulette selection and single point crossover operation of genetic algorithm.The innovation of this thesis lies in the self-adaptive particle swarm mutation rate in genetic algorithm is adapted.In the initial stage,we increase the diversity of population purposefully to avoid the premature of the system into the local optimal solution and cannot get the ideal results.While in the final stage,we reduce the mutation rate to reach the goals of maintaining the validity of the mutation,which makes sure that the mutation is carried out in an effective way.In other words,in this method we can speed up mutation to achieve the optimal solution.The combination of genetic algorithm and the particle swarm optimization algorithm is applied to the optimization of network encoding link in this thesis.The new algorithm not only eliminates the blindness of the genetic algorithm,but also makes up for the shortcomings of the premature phenomenon of particle swarm optimization algorithm.Self-adaptive particle swarm mutation rate of genetic algorithm through the simulation and comparison with the differential hybrid algorithm and the previous genetic algorithm with improved the crossover and mutation probability,the result proves that our improved algorithm is able to faster convergence speed,and can be applied to solving the problem of optimizing network coding links with better performance.