Benchmarking the restoration of heavily loaded networks using a two dimensional order-based genetic algorithm

Modern telecommunication network operators are demanding high utilisation, fast and robust networks. Network equipment manufactures are investigating techniques to design robust networks while maintaining their high utilisation. One technique is based on span/path restoration within connection-oriented mesh networks. Until now these restoration algorithms have only been evaluated in networks that include enough redundant capacity to permit 100% restoration, which is not repressive of a high utilisation network. Simulating these algorithms over a high utilisation network will not yield a possible 100% restoration and presents a problem to discover the maximum realisable restoration performance. This paper presents a method that utilises a two dimensional order-based genetic algorithm to calculate a benchmark as a realistic comparison for the different restoration algorithms.