On the performance of digital back-propagation for imperfect knowledge of link design

Digital back-propagation (DBP) was shown to have the potential to become the most promising method for compensation of fiber nonlinearities. However, apart from its high computational effort, the performance of DBP is dependent on the knowledge of the transmission link, i.e. the link´s dispersion map as well as its power profile. In this contribution, we investigate the performance of DBP under the assumption of an imperfect link-design knowledge, and discuss its implications in an optical network scenario.