A new model and algorithm for a kind of task and resource scheduling problems in elastic optical networks

Yang Qingdong,Wang Ziqing,Guo Wei,Wang Yuping

DOI: https://doi.org/10.3969/j.issn.1008-5513.2022.03.010

2022-01-01

Pure and Applied Mathematics

Abstract:Elastic optical networks using network function virtualization technology are a new type of networks. It is a challenging and key problem for simultaneously completing three tasks in elastic optimal networks, i.e., the reasonable deployment of data centers, the efficient routing and the optimal allocation of spectrums in the networks. In order to tackle this problem, a new optimization model for resource allocation and task scheduling in elastic optical networks is established first. The new model aims at most balanced load of deployment of virtual network functions on data centers, the shortest path through the assigned data centers and the fewest spectrum usage. As a result, the model can simultaneously optimize the schemes of the deployment of virtual network functions in the data centers,routing and spectrum allocation. Secondly, to solve the model effectively, we design a scheme of path selection which can find the shortest path through the assigned data centers and schemes of routing and spectrum allocation, respectively, construct initial population generation strategy and design a mutation operator. Based on these, we propose a new evolutionary algorithm for the model. Finally,we conduct the simulations, and the results indicate that the maximal used frequency slot number and the standard deviation of the numbers of VNF deployed on data centers are significantly lower than those of the compared algorithms, and also indicate the effectiveness of the proposed model and algorithm.