Research On Network Coding Links Optimization Based On The Genetic Algorithm

In traditional network transmission, nodes only storage and forwarding for information, and thus the network generally can not achieve the maximum multicast rate. Network coding raised by R.Ahlswede, is a new network transmission method, and its core idea is allowed among network nodes to mix encoding for different input link information to make the network multicast rate achieve max-flow bounds. Applied to the different areas of communication respectively, it can effectively improve the network traffic, save network bandwidth, improve network security. Network coding bring us the benefit, but also making the CPU calculation burden and cache consumption increase in the same time.Kim et al. put forward the optimization of network coding links to reduce cost. In this paper, the problem of network coding links optimization was improved. We firstly use graph decomposition method constructing the network coding optimization model, then we propose the genetic algorithm with memory function(MGA, Genetic Algorithm with Memory). This paper got the MGA Algorithm by the way that introducing crossover operator, trust, neighborhood and all one vector for the simple genetic algorithm. Simulation experiment results show that the speed of-MGA get the solution form network coding optimization model is much faster and the quality of the solution is better (that is to say the number of the coding node is less in the network coding scheme).