Optical network dynamic spectrum partition method, device, storage medium and computer equipment

Abstract

The embodiments of the present application disclose a method, device, storage medium, and computer equipment for dynamic spectrum partitioning of an optical network. The method includes: when a first traffic request is received, the The partition information and the first spectrum partition action are input into the initial spectrum partition neural network, and the score value of the first spectrum partition action is obtained; when the second traffic request is received, the second environment state and the second spectrum partition information of the elastic optical network of each sample are obtained. and a plurality of second spectrum partition actions to obtain the target spectrum partition action score value; carry out training according to the first spectrum partition action score value and the corresponding target spectrum partition action score value, and obtain the target spectrum partition neural network; obtain through the target spectrum partition neural network The maximum spectrum division action score value and the corresponding spectrum division action are adjusted together, which can improve the fairness of each spectrum division after adjustment.

Description

Technical field [0001] The present application relates to the technical field of optical network spectrum partition, and in particular to a light network dynamic spectrum partition method, a device, a storage medium, and a computer device. Background technique [0002] Elastic Optical Network (EON) is a powerful candidate for the next-generation optical communication network. In the elastic optical network, the entire spectrum is divided into a plurality of frequency gaps, typically 12.5 GHz, different services, depending on their data rate, and the connection quality allocate different quantities. [0003] However, the elastic optical network faces a wide fragmentation problem because bandwidth fragments refer to the frequency of unorganized and discontinuous frequency during the routing and spectrum allocation process, which can lead to a decline in network resource utilization and blocking. Rate increase. In order to reduce the adverse effects of fragments, scholars have propo...

AI Technical Summary

Problems solved by technology

[0005] In actual elastic optical network applications, since traffic changes with time, partition boundaries should be automatically adjusted to adapt to traffic changes. Although existing spectrum partition adjustment schemes can adaptively adjust spectrum partition boundaries according to changes in traffic requirements, However, when adjusting the partition boundaries of the spectrum, better efficiency is obtained at the cost of losing fairness, so it is impossible to balance the efficiency and fairness when adjusting the partition boundaries of the spectrum

Images