A multi-agent evolutionary algorithm based QoS unicast routing and wavelength assignment scheme

In this paper, a QoS (Quality of Service) unicast routing and wavelength assignment scheme in IP/DWDM (Dense Wavelength Division Multiplexing) optical Internet is proposed. To solve the routing problem and wavelength assignment, MEA (Multi-Agent Evolutionary Algorithm) and JFF (Joint First Fit) are used in the scheme. To realize the goal of achieving or approaching the Pareto optimum under Nash equilibrium for the provider utility and the user utility, a fair pricing method is devised by applying the principles in microeconomics and game theory. Probability theory and fuzzy mathematics are also employed to solve the uncertainty in the link state parameters and the inflexible QoS. The simulation results demonstrate good performance of the scheme.