Research On Biogeography Algorithms And Its Application In Virtual Network Mapping

With the rapid development of Internet technology growing number of users, the current network architecture already can not meet the development requirement of the new type Internet. And the network rigid phenomenon began to appear. Network virtualization provides a way to solve the problem. While the virtual network mapping problem must be solved first to realize network virtualization. When we map the virtual network to the substrate network, our main purpose to research the virtual network mapping is how to improve the utilization rate of the physical network resources. Most of the traditional virtual network mapping algorithm using two phase mapping method, the relevance between node mapping and link mapping is small, and lack the overall consideration of the entire physical network resources. Taking advantage of the characteristics of intelligent algorithm for global optimization to solve the virtual network mapping problem caused the extensive research of scholars.Biogeography-Based Optimization (BBO) is a new kind of intelligent algorithm which simulates population survival, migration and extinction rules. It has a unique migration mechanism, each individual shares its information through migration operator and don’t need to generate a new individual, they have their own mutation probability. BBO has good performance in many benchmark function tests although its structure is simple, so the prospect of researching BBO is widespread, and it can improve the ability of solving practical problems.This paper improved the BBO algorithm based on the analtsis of the standard BBO algorithm, and applying it to solve the problem of network mapping. The main research content includes:(1) On the basis of the standard BBO algorithm, introduced the cooperation mechanism, Therefore, a new Cooperative Coevolutionary Biogeography-based Optimization (CBBO) combining divide-and-conquer paradigm is proposed in this paper. In CBBO, the whole population is divided into several sub-populations and each is evolved with BBO. For each sub-population, habitats in current sub-population are constructed the context vectors with the best habitat of the other sub-populations, then calculate the fitness function value and put out the optimal solution.(2) The BBO algorithm is used to optimize the virtual network mapping problem based on the cost, a new Virtual Network Embedding Algorithm Based On Biogeography-Based Optimization (VNE-BBO) is proposed in this paper. To redefine the BBO algorithm parameters, initial population, each individual in the population corresponding a virtual network node mapping scheme. With the cost of the substrate network as the objective function, finding the optimal link mapping scheme by the iteration of BBO algorithm, and final output virtual network mapping scheme. In VNE-BBO algorithm, global searching for the optimal solution by using the intelligent optimization improved the average income of the underlying network, and reduced the costs of the underlying network overhead.(3) The BBO algorithm is used to optimize the virtual network mapping problem based on Load Balancing, a new kind of Load Balancing based on Biogeography Optimization of virtual network mapping algorithm (LB-BBO) is proposed. Network load binary model is established through the redefinition of BBO parameters, with network load as the objective function, looking for mapping scheme of load balance through the iterative optimization, the underlying network resources utilization is improved and the underlying network load balanced.