Research On Network Fault Diagnosis Based On Genetic Algorithm And Causal Model

With the rapid development of communication network,the network technology is constantly updated,the network scale is growing,the network structure is becoming more and more complex,the data running in the network is also growing,enterprises and users have higher and higher requirements for the security,stability and reliability of communication network.In such a complex network structure,network fault can not be avoided,so when network fault occurs,it is more and more important to diagnose and locate network fault quickly,solve network fault effectively and recover the normal network in time.The traditional network fault diagnosis method has a large amount of calculation and a long processing period,so it is unable to diagnose the fault accurately and quickly.Therefore,the research of intelligent,modeling and rapid network fault diagnosis method will be the main direction of future research.This paper focuses on the analysis and research of all kinds of broadband network faults in the communication access network,and proposes a broadband network fault diagnosis method based on genetic algorithm and causal model.Because the communication network fault has the characteristics of sudden,dynamic and difficult to check,it is difficult to maintain.The genetic algorithm has the characteristics of parallel search and global search,simple algorithm and strong applicability.It can reduce the amount of calculation,shorten the diagnosis time,and improve the efficiency of fault diagnosis.In this paper,genetic algorithm is introduced into the fault diagnosis of broadband access network,and the implementation of causal model of broadband network fault based on genetic algorithm is studied.The main research work and content of this paper are as follows:(1)Collect and analyze all kinds of broadband fault cases in the communication access network project,find out the causal logical relationship between the fault nodes,build a broadband network fault causal model,divide the model into four layers,from the inside to the outside are fault phenomenon layer,fault identification layer,fault cause layer and fault measure layer.(2)According to the established causal model,a mathematical model composed of objective function,decision variables and constraints is established.Taking the objective function in the mathematical model as the fitness function of genetic algorithm,the minimum value is obtained,that is,the maximum probability of path occurrence is transformed into the problem of the shortest path distance.(3)Based on the sample data of probability size between fault nodes obtained from questionnaire survey,the distance size of each path is calculated by parallel search of genetic algorithm,and the distance of each path is sorted from the best to the next,forming a relatively intuitive and extensible broadband fault path priority table.Check one by one according to the path sequence from the best to the next in the table to find out the final cause of the fault.(4)Take the sample data from the engineer's fault log in the actual communication access network project as the test set to verify,use R language to realize the effect verification of genetic algorithm,and finally evaluate the performance of the model by using the promotion degree before and after the model.The fault diagnosis model of broadband network proposed in this paper has been proved by training set and test set,and the diagnosis result is consistent with the reality,and has a high promotion effect,which shows that the diagnosis method is feasible,has a high diagnosis reliability and practical value.