Research On Access Network Planning Algorithm Based On Similarity Analysis

With the rapid development of the Internet industry,new network services emerge one after another,which brings great challenges to the carrying capacity of network equipment of the access network.As the closest end to the user side,access network has a direct impact on the user’s experience,which makes its transformation and upgrading a big difference.currently,the network upgrade planning is mainly carried out according to the experience of engineers,and the PON nodes in a certain area are reformed and upgraded through the guidance of human experience.But the upgrade process can’t be quantified by human experience,and the decision process based on index is relatively one-sided.This makes it difficult to find out which PON node need to be upgrade when the investment cost of operators is limited.The problems faced in network upgrading including(1)a comprehensive assessment scheme is lack to quantify the network PON node planning priority.(2)a network planning tools to support quantitative upgrade is lack.In view of these problems,this article focuses on the evaluation of network PON nodes,and combines the evaluation model and related manuals to design a network planning algorithm that meets the current network conditions.The main innovation and the work done is as follows:1.This thesis uses weight allocation or evaluation tools such as AHP and TOPSIS to evaluate network node priority.The human influencing factors are gradually reduced,and three evaluation models,including direct experience evaluation method,AHP+score accumulation method,and AHP+TOPSIS method,are presented to guide the network PON nodes upgrade ranking,forming a comprehensive and scientific quantitative scheme to determine the PON nodes upgrade priority under a complete data set.2.This paper evaluates,select and supplements the evaluation schemes based on data similarity.First,the variance and Jacard coefficient are used to compare and select the schemes from the aspects of stability and accuracy.Secondly,for the PON port that cannot be quantitatively evaluated with missing indicators,the calculation idea of data feature similarity under spatial coordinates of K-means algorithm is used to guide whether it should be upgraded with Euclidean distance similarity.3.According to the different operation strategies,this paper designs three kinds of upgrade algorithms for the same station’s capacity expansion in terms of practical operation simplicity,upgrade value and regional balance.These algorithms can make full use of 10G PON resources after the first time upgrade,and its planning process is based on the number of PON need to be upgraded customized by the operator.After the simulation verification of the planning algorithm,the regional equalization algorithm has the widest planning coverage and higher planning accuracy.Considering the advantages of the other two algorithms,the planning algorithm is recommended in this paper.