Research On Optimization Method Of Network Coding Based On An Improved Genetic Algorithm

In network’s multicast communication, the traditional router can only store andforward data, so data transmission rate can’t reach the maximum flow of networktransmission. Network coding technique has totally changed the situation, and through theintermediate nodes coding process, the network can reach the maximum flow of thenetwork’s data transmission.However,with the application of the network coding technique, because of extraexpense of storage and computation of data owing to coding and encoding operation, theadvantage of network code to improve the throughout can hardly be represented, so, thestudy of network coding optimization technique is inevitable.The optimization method of network coding based on the improved genetic algorithmstudies to find the minimum link cost coding scheme using network coding technique inone source multicast network and achieving the maximum multicast rate. To tackle theshortcomings which exist in general genetic algorithm such as the randomly generatedinitial population strategy can hardly make the coding data decoded so that the searchcapability of algorithm is reduced, the immature convergence problem, algorithm’s toolong operating time and so on, the improved algorithm makes some improvements such asa certain percentage of all “1” individuals are added to the randomly generated groups, soas to ensure the existence of individual which can decode the data successfully; adopt therestart solution to prevent the premature problem; narrow the size of the individualthrough not considering the link which is not able to encode data so as to improve the rateof searching, and adopt the elite individual reserved strategy to improve the convergencerate and so on. In addition, replace the roulette wheel selection of classic genetic algorithmof stochastic tournament. Proved by the simulation experiments of one source multicastnetwork coding process on VC++6.0, the improved algorithm has better performance onthe quality and efficiency of searching the optimal coding solution.