A Deep Reinforcement Learning Traffic Grooming Method in Cloud-Fog Elastic Optical Networks

Abstract

The present invention proposes a deep reinforcement learning flow grooming method in a cloud-fog elastic optical network, the steps of which are: calculating the shortest path of a service request through the shortest path algorithm; converting the service path and the network topology sliced ​​by wavelength into the form of a picture ;Use the convolutional neural network to extract the features of all pictures, use the softmax classifier to classify, and assign the service request to the corresponding wavelength; if the assigned wavelength has available resources, the service request is successfully assigned, otherwise traverse all according to the first adaptation method The wavelength allocates service requests; uses the reinforcement learning algorithm to evaluate, updates the network status of the topology, and generates the shortest path topology map for the next service request; whenever at least three service requests are allocated, the convolutional neural network is updated. The invention continuously updates the network through reinforcement learning, so that all services can make full use of ports, transceivers and amplifiers in the network, thereby reducing the total energy consumption of the network.

Description

Technical field [0001] The present invention relates to optical networks and elastic cloud - Field fog communications, particularly to a cloud - depth reinforcement learning method haze elastic traffic grooming optical network, the optical network when the resilient nodes and as a communication facility fog cloud data center, use the depth of reinforcement learning for business counseling. Background technique [0002] Cloud computing all the data to the same set of transport for analysis, storage and processing, specializes in providing a variety of services, but with the explosive growth of the global Internet of Things equipment, huge amounts of data generated by these devices are not suitable for all by cloud computing processing, and redundant transmission process will cause delay too high, given the current communication network has brought great challenges. In order to meet a lot of things low-latency computing needs, to make up for the deficiencies of traditional cloud, f...

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