This article studies the problem of constructing optimal layered multicast with network coding for heterogeneous networks. Based on the flexibility of layered source coding,a global-favorable optimization scheme is proposed,which maximizes the aggregate throughput of heterogeneous sink nodes for layered multicast with network coding by determining the optimal bit rates of the layers.To solve this global-favorable optimization scheme,especially in the large-scale heterogeneous networks,a new problem-specific genetic algorithm(GA) is further proposed.It not only searches efficiently for the optimal allocation of layer bit rates,but also guarantees the validity of candidate solutions in the whole evolutionary process.Simulation results demonstrate that this new GA-based optimization scheme could obtain efficiently the optimal or satisfactorily near-optimal bit rates for layered multicast with network coding,even in the large-scale heterogeneous networks. This article studies the problem of constructing optimal layered multicast with network coding for heterogeneous networks. Based on the flexibility of layered source coding, a global-favorable optimization scheme is proposed, which maximizes the aggregate throughput of heterogeneous sink nodes for layered multicast with network coding by determining the optimal bit rates of the layers. To solve this global-favorable optimization scheme, especially in the large-scale heterogeneous networks, a new problem-specific genetic algorithm (GA) is further proposed. Not only searched efficiently for the optimal allocation of layer bit rates, but also guarantees the validity of candidate solutions in the whole evolutionary process. Simulation results demonstrate that this new GA-based optimization scheme could be as good as the optimal or satisfactorily near-optimal bit rates for layered multicast with network coding, even in the large-scale heterogeneous networks