2020 · 23 citations · 10 references
ComputingComputer EngineeringComputer ScienceComputer NetworksDesign OptimizationOptimizationEvolutionary OptimizationGenetic AlgorithmsLatency OptimizationMetaheuristic OptimizationNetwork OptimizationSoftware-defined Networking5G NetworksClusteringComputational Complexity Of AlgorithmsGateway PlacementMetaheuristic SearchNetwork Performance OptimizationPlacement AlgorithmsSimulated Annealing
Software-Defined Networking paradigm has been proposed for the integrated 5G-satellite network to achieve better flexibility, scalability and reliability. A particularly important task in such the SDN-enabled unified architecture is that of the joint gateway and controller placement problem to meet various requirements, especially for the network reliability and latency. In general, an exhaustive evaluation of all placements can produce optimal results but having exceptionally long running time. In this paper, we proposed two metaheuristic approaches that can achieve approximate optimal performance with much lower computational complexity. A Simulated Annealing based algorithm called Double Simulated Annealing is proposed. Further, a Genetic Algorithm based method was developed to achieve the near-optimal of both controller and gateway placements simultaneously. Comprehensive experiments were carried out to select the best hyper-parameter settings for both algorithms. Simulation results demonstrated that the Double Simulated Annealing algorithm outperformed the state-of-the-art Simulated Annealing and Clustering Hybrid Algorithm in terms of solution accuracy with a small trade-off of running time. Moreover, the Genetic Algorithm based approach can provide a near-optimal solution with better precision and lower computational complexity in comparison with the Simulated Annealing and Clustering Hybrid Algorithm.
10
Simon Knight, Hung Nguyen, Nickolas Falkner et al. · IEEE Journal on Selected Areas in Communications · 2011 · 1.7K citations