هوش محاسباتی در تکنولوژی فوتونیک و شبکه های نوری: بررسی و دیدگاه های آینده

The continuous growth of broadband communications, multimedia services and Internet is absolutely related to the deployment and operation of optical networks. Despite optical fibers’ enormous physical bandwidth the development of optical networks for today’s advanced, reliable and guaranteed-type services, require an efficient management of the bandwidth together with an orthological and careful use of optical components given their high manufacturing cost. These requirements have lead to the need for sophisticated photonic devices and to optical networks’ implementations of increased functionality and associated thus complexity. For the efficient consideration of those problems different design and optimization techniques have been applied to date. However, as the complexity increases, the use of computational intelligence (CI) in those problems is becoming a unique tool of imperative value. In this paper we review in a unified approach the applications of CI starting from the physical layer and ending to services layer, given that here there is a strong relation and unique interplay between components’ technology and network issues, being sharing the common target of physical bandwidth’s efficient utilization. The applicability of different CI classes (genetic algorithms and evolution strategies, fuzzy systems, and artificial neural networks) in optical wavelength division multiplexing (WDM) networks is identified and evaluated. Furthermore specific optical networks’ optimization problems are categorized. Being a rapidly growing area, new trends, such as evolutionary game theory, in understanding and design of large scale Optical Network are also identified and discussed. The paper seeks to review the aforementioned areas, identify new problems and trends, triggering this way new research efforts for interdisciplinary cooperation between researchers.