Sizing Eligible Route Sets for Restorable Network Design and Optimization

In the use of linear programming models for optical transport network design and optimization, appropriately selecting the number of eligible restoration routes is an important factor in the tradeoff between runtime and optimality. This is because too large an eligible route set results in excessive runtime, while too small an eligible route set results in design solutions that could be significantly sub-optimal. We analyse this tradeoff on a wide range and variety of test networks in span- restorable, path-restorable and p-cycle networks and discuss how our findings can be very useful in determining the appropriate number of eligible route set sizes to use for a particular network design and optimization. We also validate these results with a follow-up set of test networks.