Multi-objective Network Coding Optimization Based on NSGA-II Algorithm

Abstract

Network coding could effectively improve transmission performance of multicast network, but encoding of node brings the additional calculation cost of node. In order to overcome the overhead brought by network coding, a network coding optimization model under the framework of algebraic network coding is designed in this paper, and the joint optimization for both the link cost and coding cost of network coding are carried out based on this model. In addition, the paper proposes MOONC (Multi-Objective Optimization Problem of Network Coding) based on improved NSGA-II algorithm. Adopting non-dominated sorting mechanism, virtual fitness and elitist strategy, this algorithm could not only improve algorithm efficiency and convergence speed, but also guarantee the population diversity. The simulation for typical network topology shows that this algorithm is effective and feasible.