Optimization of Resilient Networks with Column Generation

An optimal cost-reducing capacity-dimensioning is crucial for network operators when designing communication networks. Failure-free working as well as backup routes have to be installed in order to guarantee Quality of Service and availability requirements. With traditional network planning techniques using linear programming with the flow-approach [1] or the path-approach [2] optimal configurations can be calculated for small-size networks only. In this paper we present a network planning approach based on linear programming with cut-and-price for the optimization of path protection and path restoration. With this approach paths and path-variables are generated during the optimization process. The memory consumption and equation complexity can be reduced while optimality of solutions can still be guaranteed. The new optimization approach is capable of calculating optimal configurations for medium and large size networks and reduces calculation times and memory consumption.