Research On Network Link Selection And Optimization

With the rapid growth of Internet consumer groups,the network data shows explosive growth,the network scale is expanding,and the network topology is becoming increasingly complex.At the same time,the real-time requirements for network services are constantly improving,which requires the system to provide a fast algorithm to calculate the shortest path between nodes when routing;There will be a large number of network request services in the same time in the computer network.In order to prevent the network link from overload or even network congestion,it is required that the system can improve an excellent network optimization algorithm to distribute the services reasonably and improve the overall utilization of network resources.Starting from the two directions of the shortest path selection and the whole network optimization,In the aspect of shortest path algorithm,we propose an improved ALT algorithm;In the aspect of the whole network link load optimization,we propose an improved Gray Wolf Optimization algorithm.Finally,we apply the improved ALT algorithm and the improved Gray Wolf Optimization algorithm to the system.The main contents of this paper are as follows:1.In view of the low efficiency of ALT algorithm,we propose an improved algorithm abt algorithm,which improves the distance prediction function.In addition,in terms of the selection strategy of backbone nodes,we propose a greedy strategy based backbone node selection algorithm cgbs algorithm.We have proved by experiments that the performance of ABT algorithm is better than that of ALT algorithm in the online route finding stage,and ABT algorithm is suitable for large-scale network.2.Aiming at the problems of low convergence accuracy and easy to fall into local extremum of Gray Wolf algorithm,we propose an improved Gray Wolf algorithm,which is based on elite reverse learning strategy and the best position strategy of individual history.Its main idea is: on the basis of the current optimization solution set,according to the elite reverse learning strategy,the algorithm searches for the corresponding reverse elite solution to enrich the diversity of the population;In order to truly reflect the convergence process of the algorithm in engineering application,the algorithm adopts an adaptive variable convergence factor based on cosine function;Based on the basic principle of Particle Swarm Optimization,the algorithm integrates the best position of gray wolf in history.We prove the effectiveness of the algorithm through simulation experiments,and apply the algorithm to the network load optimization model.3.We design and implement the network link selection and optimization system based on the improved ALT algorithm and Gray Wolf algorithm.Then we introduce and show the design idea and implementation process of each module in detail.The system can monitor links and services in real time.Finally,we carry out the test of route selection and load balance to test and verify the effectiveness of the system.