Minimum CAPEX/OPEX design of optical backbone networks

Abstract

We develop new scalable tools for optimizing the design of agile optical backbone networks while minimizing capital and operational expenses (CAPEX/OPEX). The proposed tools not only aim at the dimensioning of the available equipment (e.g., photonic cross connects - PXCs), but also address the cost issue throughout the choice of the cheapest node switching equipment for a given grade of service. For this reason, we propose an original optimization model where we minimize the number of required switching fabrics, determine their optimal node location and dimensioning, while granting all traffic requests. We next suggest an efficient large scale optimization solution scheme, based on column generation techniques, in order to solve the resulting model, so that, in spite of its size and combinatorial features, we are able to obtain its optimal solution in a reasonable amount of time. Experiments are conducted on different network and traffic instances. Results show that it is not necessary to provide PXC equipment at every node in a network in order to reach a minimum CAPEX/OPEX network design, only few of them in well selected nodes are sufficient in order to satisfy all demand requests, at minimum cost. The optimal network designs not only minimize CAPEX/OPEX, but also provide solutions with the minimum number of nodes requiring a PXC equipment.