Optical packet switching over arbitrary physical topologies using the Manhattan street network: an evolutionary approach

Published in "Towards an Optical Internet", A. Jukan (Ed.). Optical packet switching over arbitrary physical topologies typically mandates complex routing schemes and the use of buffers to resolve the likely contentions. However, the relatively immature nature of optical logic devices and the limitations with optical buffering provide significant incentive to reduce the routing complexity and avoid optical domain contentions. This paper examines how the Manhattan Street Network (MSN) and a particular routing scheme may be used to facilitate optical packet switching over arbitrary physical topologies. A novel approach, genetic algorithms (GA), is applied to the problem of deploying the MSN (near) optimally in arbitrary physical topologies. A problem encoding is proposed and different implementations of GA described. The optimum GA parameters are empirically selected and GA is successfully used to deploy the MSN in physical topologies of up to 100 nodes. Favourable results are obtained. GA are also seen to out-perform other heuristics at deploying the MSN in arbitrary physical topologies for optical packet switching