Abstract

First and foremost, I promise a (reasonably) light talk, with limit equations and multiple (multicoloured!) pictures—sharks may even feature, if we have time!

In this talk, I shall discuss challenges in ‘routing schemes’ for fibreoptic networks, and suggest a possible scheme. In essence, one (eg BT or Virgin Media) has a network of cables connecting cities, and wishes to send data down these cables; however, one cannot send two signals of the same type down the same cable, as they then interfere. I shall formulate this precisely, and it turns out that finding a (suitable) routing scheme is an NP-hard problem.

Using randomness (Markov chains), I shall show how one can find a suitable solution… provided one’s definition of “suitable” is, well… suitable. These types of systems are actually used in the real world.

For those who, like me, do pure maths, I shall explain this (likely previously unknown) notion of the ‘real world’