Design And Simulated Implementation Of Game Theory Based Energy-saving Grooming Algorithm In IP Over WDM Optical Networks

With the popularization of the Internet and the rapid development of network technology, IP traffic presents the explosive growth and the traffic granularity also shows to be diversified, which urgently needs the optical network to be able to assign network bandwidth dynamically to meet traffic requirements. Since the bandwidth of a connection request is generally much less than that of a full wavelength, if a wavelength is dedicatedly assigned to each connection request, a huge waste of network resources will be produced. Therefore, the traffic grooming technology has been emerged. Meanwhile, the continuing expansion of network makes the energy consumption increase, which not only limits the development of network but also be out of tune with the current concept of energy conservation and low-carbon life. Therefore, the concept of building green networks is put forward in recent years.Based on the above ideas, the traffic grooming problem which can promote the energy-saving is studied in this thesis. Firstly, the node structure with the optical bypass function and the network model is built, and the power consumption statistical model is proposed. Based on these models, two kinds of energy saving grooming mechanisms including the single-hop energy-saving grooming mechanism and the multi-hop energy-saving grooming mechanism are designed. Considering the interest conflict between users and network operators, the game theory in microeconomics is introduced into solve the non-cooperative interest relationship between them and to achieve the "double-win" solution. Furthermore, a fitness function which can evaluate the power consumption and the benefit of users and network operators on the paths is proposed, and two energy-saving grooming algorithms including Single-hop Energy-saving Grooming algorithm based on Game theory (SEGG) and Multi-hop Energy-saving Grooming algorithm based on Game theory (MEGG) are designed to help each connection request to select the path with the optimal integrated evaluation (fitness value).In order to evaluate the performances, the platform of VC++6.0is used to simulate the designed algorithms. Simulation results show that the algorithms designed can obviously reduce the mean power consumption and the blocking probability, and it also can effectively balance the benefit between users and network operators to achieve the "double-win".