Deep reinforcement learning traffic grooming method in cloud-fog elastic optical network

Abstract

The invention provides a deep reinforcement learning traffic grooming method in a cloud-fog elastic optical network. The method comprises the steps: calculating a shortest path of a service request through a shortest path algorithm; converting a service path and the network topology sliced according to the wavelength into a picture form; extracting features of all the pictures through a convolutional neural network, carrying out the classification through a softmax classifier, and distributing the service requests to corresponding wavelengths; if the allocated wavelengths have available resources, successfully allocating the service requests, otherwise, traversing all the wavelengths according to a first adaptation method to allocate the service requests; evaluating by using a reinforcement learning algorithm, updating the network state of the topology, and generating a shortest path topological graph of the next service request; and updating the convolutional neural network after theallocation of the at least three service requests is completed each time. According to the method, the network is continuously updated through reinforcement learning, so all services can fully utilizeports, transceivers and amplifiers in the network, and the total energy consumption of the network is reduced.

Description

technical field [0001] The present invention relates to the technical field of elastic optical network and cloud-fog communication, and in particular to a deep reinforcement learning traffic grooming method in cloud-fog elastic optical network. When the elastic optical network is used as a communication facility for fog nodes and cloud data centers, Business grooming using deep reinforcement learning. Background technique [0002] Cloud computing transports all data sets to the same center for analysis, storage and processing, and is good at providing various services. However, with the explosive growth of global Internet of Things devices, the massive data generated by these devices is not suitable for all through cloud computing. At the same time, the redundant transmission process will cause excessive delay, which brings huge challenges to the current communication network. In order to meet the needs of a large number of low-latency computing in the Internet of Things an...

AI Technical Summary

Problems solved by technology

[0015] In the existing research, only fixed traffic grooming strategies or simple strategies relying on manual feature extraction cannot achieve true adaptive traffic grooming strategies

Images