Abstract
Outage probabilities due to polarization-mode dispersion (PMD) are typically required to be very small. This constraint makes it extremely difficult to use either Monte-Carlo simulations or laboratory experiments to determine the outage probability of a system, because of the extremely large number of configurations that must be explored in order to obtain reliable estimates. The rare events where PMD is unusually large are the ones most responsible for generating system outages. A natural measure of PMD is the PMD vector,1 the magnitude of which is the differential group delay (DGD). Recently,2 we have applied importance sampling (IS)3 to numerical simulations of PMD. In this method, the IS works by biasing the Monte-Carlo simulations so that the configurations producing large DGD events are realized much more frequently. The method makes it possible to directly calculate the outage probability of systems for which first-order PMD is of primary significance.4
© 2002 Optical Society of America
PDF ArticleMore Like This
W. Kath and G. Biondini
ThA1 Optical Fiber Communication Conference (OFC) 2003
Sarah L. Fogal, Gino Biondini, and William L. Kath
ThA2 Optical Fiber Communication Conference (OFC) 2002
A. Lima, I. Lima, B. Marks, C. Menyuk, and W. Kath
ThA3 Optical Fiber Communication Conference (OFC) 2003