Optical Transport Network Design via Distributionally Robust Nonlinear Optimization

Optical transport network (OTN) design, as an emerging technology required for the development of next-generation networks, is often filled with complex uncertainty data failing to satisfy the classical i.i.d. conditions. To address such challenges, under the assumption of nonlinear i.i.d., this article proposes a novel distributionally robust nonlinear optimization (DRNO) approach for designing OTNs. It integrates the location-routing-coverage decisions with given household population information and capacity limitations. Importantly, the proposed approach converts the underlying model equivalently into a mixed-integer second-order cone programming, by exploiting a specific nonlinear normal distribution ambiguity set and the max-mean algorithm. To further achieve the solvability of large-scale models, this article introduces two matheuristic algorithms (MAs) by analyzing the characteristic structures of OTNs. The effectiveness of the proposed approach and algorithms is verified by implementing numerical experiments based on nominal data acquired from industry.