Traffic Grooming Of Multicast WDM Mesh Networks With GA

With the progress of optical communication technology and the development of optical networks, the number of Internet users as well as the amount of traffic in the network increase exponentially. Multicast will become one of the most important communication formats in the future communication systems. These multicast messages are used by groups of end-users to provide various conveniences for their modern life. As far as we know, mesh network is one of the most widely used network formats. Because of the great increase of information, without proper optimization, the network's resource cannot be efficiently utilized, and the cost of networks would increase accordingly. For these reasons, traffic grooming becomes the key issue in constructing the high capacity and high-speed networks, which has attracted more and more attention from the academia and the industry society alike. Based on these considerations, we commence our study on the traffic grooming in multicast mesh networks.The research of traffic grooming in mesh networks is the problem of optimization of routing and wavelength assignment (RWA) (which can also be referred to as the problem of multiplexing various sessions in this paper). But the RWA in mesh networks is a NP-hard problem which can not be solved optimally with exhaustive search in the endurable time. Hence, we designed a heuristic algorithm to solve it.But the heuristic algorithm has some limitations when dealing with large scale networks. By comparing all kinds of intelligent algorithms and reviewing the literatures in traffic grooming, we realized that genetic algorithm (GA) is the edge tool for solving NP-hard problems. So we designed a unique GA approach to solve the problem of traffic grooming in multicast mesh networks.After carefully studying the way of multicast traffic propagation in the mesh networks, we utilized our GA approach to search for the optimal solutions of wavelength assignment and traffic grooming in the network. Computer simulations in non-uniform static multicast traffic in mesh network showed that our GA approach could achieve satisfied grooming results.These results can provide useful reference to the optimal design of WDM multicast mesh networks.