| We investigate the problem of multicast routing and wavelength assignment under the constraint of sparse light splitting or power budget.; The high cost of cross-connects capable of light splitting motivates us to investigate the problem of multicast routing in sparse splitting networks, where only a subset of cross-connects are capable of splitting light. We develop a rerouting algorithm that generates feasible solutions in sparse light-splitting optical networks. In addition, to further minimize the fiber usage in a solution, we present a heuristic based on Tabu Search that obtains solution within 10% of the optimal in almost all the instances and within 5% of the optimal in about half of the instances in our experiments.; Multi-drop path model presents a cost-effective and practical method for realizing multicast connections. In a multi-drop path model, multiple paths are employed to establish a multicast session, and each path can drop signals at no more than a predefined number of destination nodes. For the single session case, experimental results show that the proposed MFBH heuristic achieves the lower bound on the number of wavelengths in almost all instances for real-life and randomly generated networks. MFBH effectively minimizes the number of wavelengths used for a single multicast session, however, it does not perform well in minimizing the number of wavelengths used for multiple multicast sessions, because it does not consider the routing interactions among sessions.; To solve the problem of multicast routing and wavelength assignment under the multi-drop path model for multiple sessions case, a simple approach is to combine MFBH and the SL wavelength assignment heuristic proposed in literature. This algorithm is called MFBH-SL. We develop a more effective routing and wavelength assignment algorithm called TSH, which outperforms MFBH-SL in most instances for real-life and randomly generated networks. Furthermore, the wavelength saving is significant when session size is small (≤40% of network nodes). |
