Research On Dynamic Routing And Wavelength Assignment Based Heuristic Algorithm In Intelligent Optical Networks

With the rapid growth of IP services, its inherent burst and real-time characteristics have beenput forward higher requirements for current optical networks, in which Dynamic Routing andWavelength Assignment (DRWA) become a critical issue in future Intelligent Optical Networks(ION). This thesis gives a comprehensive research on DRWA issue in ION and proposes a novelDRWA algorithm with theoretical analysis and numerical simulation methods.Firstly, this thesis presents some principles and technical characteristics of ION, focusing onexisting DRWA solutions and routing issues in ION. An improved Ant Colony Optimization (ACO)based algorithm (Improved Super ACO algorithm for DRWA, IS-DRWA) is proposed. Theimproved algorithm improved node selection rules and enhanced the ability to balance convergencespeed and global search, by which the improved algorithm can avoid the local optimal solution. Theimproved algorithm employs varying link idle rate instead of fixed local pheromone evaporationfactor to make the routing which having more wavelength numbers allocate more business. Foravoiding algorithm precocity and trapping into local optimal, an adjustable global pheromoneevaporation factor according iterative results is introduced. This thesis set up different networktopology simulation environment platform and evaluated the performance of IS-DRWA algorithm.Simulation show the improved algorithm can decrease blocking probability and promote resourceutilization comparing to Dijkstra+FF algorithm.