Cross-domain virtual network mapping method based on hybrid heuristic algorithm

Abstract

The invention provides a method for virtual network mapping. According to the method, a hybrid heuristic algorithm is utilized, the problems existing in a traditional mapping method are optimized, the flexibility of the method is improved by using a dynamically calculated crossover probability and a pheromone-based mutant gene selection strategy, and a weight updating mechanism based on load balance is introduced, so that the probability of mapping failure is reduced while the load is balanced. According to the technical scheme provided by the invention, the defects of a heuristic mapping method can be made up, the mapping cost is reduced, the mapping success rate is improved, and a plurality of performance indexes such as the mapping average quotation, the link load balance, the mapping cost-benefit ratio, the acceptance rate and the operation time are good.

Description

technical field [0001] The invention relates to a cross-domain virtual network mapping method based on a hybrid heuristic algorithm, in particular to a virtual link mapping method considering load balancing. Background technique [0002] The Internet of Things, big data, cloud computing, and edge computing, as representative technologies of the third information revolution, have gradually become an indispensable part of our lives through their coordinated development. However, the coordinated development of the three technologies has led to an unprecedented increase in the number and scale of network services, which has greatly increased the demand for high-traffic network communications and flexible network construction. In this context, the traditional network architecture (that is, every adjustment needs to rebuild the underlying physical network structure) can no longer meet the demand, which leads to the problem of Internet rigidity. In this regard, a lot of attention ...

AI Technical Summary

Problems solved by technology

[0010] (1) When there are many algorithm parameters, parameter tuning is a complex and repetitive process. If you only rely on manual settings, it will not only rely too much on experience, but also lead to poor generalization ability of the algorithm and cannot dynamically adapt to various environments.

[0011] (2) The global search process guided by the optimal solution will easily make the algorithm fall into local optimum, which will affect the performance of the mapping scheme generated by the algorithm

[0012] (3) When solving special optimization problems, the search ability is too low, resulting in premature convergence

Therefore, this will inevitably affect the search ability, which in turn will lead to premature convergence

[0013] (4) For the virtual network mapping problem in discrete space, the optimization process guided by the direction vector is usually ineffective, which makes many excellent heuristic algorithms not applicable to the virtual network mapping problem

Images