An algorithm for the routing and wavelength assignment problem in WDM networks

WDM provides the ability to divide the huge optical bandwidth available in fiber into lower-capacity wavelengths, and they can economically satisfy the growing demand for bandwidth. This paper studies the RWA optimization problem in a WDM network and provides an effective approach with reasonable computational complexity and good performance to solve this problem. We formulate the RWA problem as a minimizing-rejection-penalty problem. The approach combines the Lagrangean relaxation technique, the MSPAWG, a heuristic and subgradient method. In the NSFNET example, the influence on the overall rejection penalty from the number of converters and number of wavelengths installed is discussed. Also, the trade-off between the number of rejected connection demands and the disconnection ratio, which means the fairness of the rejection decision, by using different rejection penalties, is studied.