We implement the approximation algorithm for the multicast congestion problem in communication networks in [14] based on the fast approximation algorithm for packing problems in [13]. We use an approximate minimum Steiner tree solver as an oracle in our implementation. Furthermore, we design some heuristics for our implementation such that both the quality of solution and the running time are improved significantly, while the correctness of the solution is preserved. We also present brief analysis of these heuristics. Numerical results are reported for large scale instances. We show that our implementation results are much better than the results of a theoretically good algorithm in [10].
展开▼
