Evolutionary Computation Based Performance Enhancement in WDM Networks using Wavelength Converters

Abstract

Usage of optical networks is increasing day by day to transmit the data from one place to newlineother place, one device to other device or one person to other person. Distance between newlinethe devices is very high when they are located across oceans or countries or continents. In newlineorder to increase the speed of transmission and to reduce the losses during the newlinetransmission, the electrical signals are converted into light pulses and it is transmitted newlinethrough optical fibers. The network formed by optic fiber links is known as optical fiber newlinenetworks or simply optical networks. The optical networks are of many types and some of newlinethe examples include wide area networks (WAN) or local area networks (LAN). newlineWhen a signal is to be transmitted from source node to destination node in a network, it newlinehas to be propagated through all the intermediate links between source and destination newlinenodes. Also, there are many channels or wavelengths that an optical fiber can support newlineduring the transmission. A single physical link can transmit several light pulse signals newlinehaving distinct wavelengths. There will be no overlap of wavelengths of signals when it is newlinebeing transmitted in the optical network. The main challenge is to determine the newlinewavelength for a call when it arrives for propagation through the network since there are newlineseveral channels available for assignment in the network. When a channel is assigned to a newlinecall and when it is progressing through the network, if another call arrives, then another newlinechannel is assigned. In this process, there is a possibility that there will be no free newlinechannels available for a new call at some point of time or the next link in the network is newlinealready occupied during the propagation. In such a case, the call gets blocked and the newlineperformance of the network is measured with a metric known as blocking probability. newlineIn this work, two baseline methods, namely, First fit assignment and Random newlineassignment are modeled and implemented in MATLAB for a tandem network as baseline newlinemodels. The baseline models a