An investigation into routing algorithms for optical networks

The backbone of the internet is now an optical network. It no longer has to rely on more costly electronic-optical conversion which typically led to bottlenecks at the nodes. However, the drawback of optical networks is that they rely on wavelength-division multiplexing transmission in which a common wavelength needs to be assured on all links of a route for the connection not to be blocked. Efficient use of the network is therefore highly dependent upon good wavelength-assignment and routing algorithms. An Optimisation technique, of Integer Linear Programming, has already proved its worth in this context. However, it requires a fixed set of connections and is too slow to adapt to more a dynamic market. Moreover, as new technologies emerge, the such as optical switching, and demand becomes more dynamic, for example with the introduction of Broadband-on-demand, alternative approaches will be required. Mathematical expertise will be needed to derive suitable mathematical formulations and to develop optimisation wavelength assignment and routing algorithms.By providing the opportunity for someone with broad experience of optimisation methods in other contexts, to familiarise themselves with this topic, we hope to derive some novel approaches. Inspiration may be drawn from optimisation approaches in other domains, such as Graph Colouring, and gaps in solution method will provide further challenges to the Mathematics community. In addition, the collaboration built up by working together should enable us to develop suitable routing methods in anticipation of emerging technologies. It is hoped that this will provide UK universities and industry with a competitive edge from changes in technology and modes of use of optical networks.