Degree-Constrained Multicasting for Multimedia Communications
暂无分享,去创建一个
With the proliferation of multimedia group applications, multicasting is becoming increasingly important. Multicast routing is establishing a multicast tree that is rooted from the source node and contains all the multicast destinations, which is often modeled as the NP Complete Steiner problem in networks.Finding a multicast tree is complicated by the likely heterogeneous nature of the multicast environment. Nodes in the network will likely vary in their abilities to support multicasting. Some nodes may not support multicasting; others may be limited in the number of multicast copies they can reasonably make in order to ensure network speed and distribute the load more evenly among the nodes in the network. The multicast capability of each node is represented in this paper by a degree constraint. The problem of how to create the multicast tree when nodes have different multicast capability is referred to as the degree constrained multicast tree problem, which is also a NP Complete problem. In this paper, a two layer genetic algorithm (GA) for solving the degree constrained multicasting problem is proposed. The basic idea of this algorithm is that the optimal multicast tree must be a minimal spanning tree satisfying the degree constraint, so the key to the final result is to find the Steiner nodes which should be included in the final tree. The algorithm adopts binary coding. The inner layer GA builds the degree constrained minimal spanning tree; the outer layer GA performs global exploration. The paper presents the simulation results of the algorithm proposed on sparse networks of nodes with varying multicast capability, and the simulated multicast groups are small relative to the size of the network, reflecting likely multicast applications. We experiment on the GA in three aspects: (1) quality of solution; (2) computational time; (3) convergence. Experimental results show that the algorithm proposed can find the tree with minimal cost. But the GA proposed requires much time to find a multicast tree when the network grows larger.