Virtual Network Mapping Algorithm Based On Cellular Genetic Mechanism

Network virtualization provides a new way to solve the rigid problem of traditional network structure.The key problem is how to map the virtual network to the physical network.Virtual network request optimal mapping problem which satisfies node and link constraints is mainly solved by heuristic algorithm such as particle swarm optimization and genetic algorithm.Most of these heuristic algorithms solve problems from the perspective of mathematical model optimization,departed from the characteristics of the virtual network mapping problem itself,and in the process of establishing the model,the influence of the change of the network node itself on the optimal solution is not considered.In view of the shortcomings of traditional heuristic algorithm and the change of nodes and link resources in the process of virtual network mapping,cellular automata is used to model the virtual network mapping problem,and an intelligent algorithm for virtual network mapping is designed by combining the genetic mechanism.According to the characteristics of the virtual network embedding problem,this paper uses the evolution rules of cellular automata to replace the crossover operations of traditional genetic algorithms,which makes the interaction between the individual all-round and dynamic.The neighboring structure and evolution rules are suitable for the specific mapping environment,which ensures the diversity of the population.Then the paper proposes a virtual network embedding algorithm based on cell genetic algorithm(Virtual Network Embedding-Cellular Genetic Algorithm,VNE-CGA).As the experimental results show,compared with existing algorithms,the algorithm improves the operating benefits of the underlying network and the virtual network request acceptance rate.Counter the limitations of VNE-CGA in solving virtual network embedding problems,this paper uses cellular automata to remodel virtual network embedding problems and Individual cells as individuals in the population,realize the selection and cross operation in the genetic algorithm by means of cell learning from neighbors,give the characteristics of "intelligence" to individual cells,then propose a virtual network embedding algorithm(Virtual Network Embedding-Neighbor Learning-Genetic Algorithm,VNE-NL-GA)that combines neighbor learning and genetic mechanisms.The experimental results show that,compared with VNE-CGA algorithm,the algorithm is more adaptable in different embedding environments,and it greatly reduces the running time of the algorithm.