Method and system for selecting a minimum load router based on naive Bayes classifier

Abstract

The invention relates to a method and system for selecting a minimum load router based on a naive Bayes classifier. The performance of the method for selecting the minimum load router is improved, a network snapshot records historical network state information, and the naive Bayes classifier is used for predicting the potential network blocking rate of all candidate routers between each node pair.The network snapshot corresponds to each arrived service request and records the number of occupied network resource units on each link. If the current service request establishes a connection on a candidate router, then the naive Bayes classifier predicts the potential blocking rate for future service connection establishment and calculates the load for each candidate router at the same time. Finally, the router with the smallest load and the lowest potential blocking rate and without service connection establishment is selected from all the candidate routers between the node pairs to establish service connection to achieve the optimal balance between the load and the blocking rate.

Description

technical field [0001] The invention specifically relates to a minimum load routing selection method and system based on a naive Bayesian classifier. Background technique [0002] Routing algorithms in circuit switched networks is a mature subject. Routing algorithms can be divided into three categories: fixed shortest routing algorithms, fixed alternative routing algorithms and adaptive routing algorithms. [0003] Fixed shortest route algorithm: Find a fixed shortest route for a node pair in advance, and all services between node pairs will be established along this route. [0004] Algorithm for fixed alternative routing: select multiple fixed routes between node pairs, and then try to establish services in turn, that is, the first route is preferred to establish services, and if unsuccessful, try the second route until all routes are used. Try to build a business. [0005] Adaptive routing algorithm: According to the real-time link status in the network, the optimal ro...

AI Technical Summary

Problems solved by technology

Although machine learning techniques have been widely used in various fields including communication networks, no machine learning techniques have been used to improve the performance of routing algorithms in circuit-switched networks.

Images