Differential Evolution Algorithm And Its Application Research On Network Coding

With the popularization of computers,human have now entered to information era,and the network techniques have been applied in many fields.People are more and more depending on the information network.However,the transmission efficiency of the current networks is still limited,which can't satisfy the users' increasing requirements of high transmission speed on networks.Network encoding technology appears to ease this problem.Network coding makes each node in the network recode the received data before transmission rather than simply forwarding.Network coding reduces the number of data transmission and increases the amount of information in a single transmission.It also improves the network throughput,robustness and security.In other hand,nodes need to additional encoding operations which increase the network complexity and overhead.How to find the optimal network coding strategy is particularly important.In this paper,we propose a network coding optimization method based on simulated annealing,differential evolution and existing network coding techniques.First,we introduce the idea of simulated annealing and individual acceptance mechanism into differential evolution.It ensures the population diversity and convergence of groups in the evolution,and improves the global optimum locating ability of network coding,which is called SDE.Secondly,an adaptive differential evolution algorithm,SVDE,based on simulated annealing is proposed.In this algorithm,the scaling factor in variation process of differential evolution is dynamic.In order to find the optimal solution to increase the diversity of population,we expect the differential of groups decline in later period of iterations.We introduce a linear transformation into scaling factor to ensure enough diversity of groups in early and convergence as soon as possible in later.At last,simulating experiments on different kinds of butterfly network graphs show that the proposed algorithm,SDE and SVDE,can find less number of edges in the network coding scheme in less time,when compares with conventional differential evolution algorithm.It indicates that the performance of SVDE algorithm is better than that of SDE algorithm.