Utilization-Based Optimal Placement Of Wavelength Converters In WDM Optical Network Using New Mutation Strategy In The Differential Evolution

Abstract

In this paper, a new mutation strategy in the Differential Evolution has proposed for an optimal
allocation of wavelength converters in the wavelength division multiplexing (WDM) optical
network. Concerned objectives were to maximize the overall utilization performance of placed
wavelength convertors along with handling the network conflict conditions of placement
efficiently. Traffic handling by the different number of wavelength converters has considered in
extracting the overall performance of converters placement. A better traffic handling model has
also proposed as an objective function for efficient utilization of allocated convertors when there
were multiple equivalent conditions of traffic load existed in the different parts of the network. To
achieve the objectives, under the Differential Evolution a new mutation operator and formation
of the efficient differential vector have proposed for faster convergence and better exploration
for the solution. Convergence rate has improved by obtaining the differential vector under the
guidance of the best solution or through the randomly selected solution from the population in a
probabilistic way. The level of exploration has increased in the process of mutation vector
formation by applying the logical XOR operation between the differential vector and the base
vector. The proposed approach has applied over benchmark network model NSFNET and it has
delivered a more effective performance in comparison to other forms of Differential Evolution as
well as reasoning based results available in the past literature.