The Application Of Artificial Immune Algorithm In Communication Network Fault Diagnosis

With the widespread adoption of comp uter equipment and the development of network technology, communication network presents the development of complicated structures, diverse equipment, scalable business. Modern communication network is no longer a single traditional hierarchy, voice, image, video and other multimedia business have also join them. Therefore, when the network fails, it can produce a large number of alarm, multiple alarm event is overlay, real useful alarm information is submerged, thus bringing a lot of trouble to the fault diagnosis. And the fault diagnosis technology is to find the root of the problem that is the fault from a mass of the presentative alarm information, so how to complete fault diagnosis in a large scale and complex network environment accurately and efficiently becomes a major of the research hotspots.In the running process of communication network, the performance indexes of the network equipment are dynamically changing, thus we can "sense" the network condition by monitoring these data. The artificial immune algorithm references the biological immune principle, which has the characteristics of self- learning, self-adaption, quick response. In this paper, we use the characteristics of artificial immune algorithm and combined with the decision tree classifier to put forward an artificial immune synthesis fault diagnosis method, and apply it to the fault diagnosis of communication network. The specific process mainly includes: first, get the normal operation data when communication network operates normally, and get the fault data when it fails, use the dynamic threshold negative selection algorithm to train the normal operation data and obtain the anomaly detector, use the decision tree combination method to train the fault data and obtain the decision tree combination classifier; Then, collect the test data from communication network for a period of time, use the anomaly detector to analyse the test data: if the test data exists abnormal, it will be diagnosed, otherwise the data is normal; Finally, use the decision tree combination classifier to diagnose the abnormal data, and determine its fault type.The artificial immune synthesis fault diagnosis method combines the characteristics of artificial immune algorithm and decision tree classifier, in the aspect of anomaly detection it can diagnose the "unknown" abnormal, and the "non self" spatial coverage has been greatly improved; in the aspect of fault type diagnosis, the fault can be found from the abnormal data and its type can be determined, and the classification accuracy of the decision tree is greatly improved.