Immune Algorithm Research For Network Fault Diagnosis

As the network have a wide range of applications on people's work and live, the importance of network fault management is significant increasingly. The expansion size and complex structure of the network make the network management and maintenance become more difficult. There are many factors that caused the fault in the network, the occurrence of fault is inevitable, discovering the network fault timely and determining the position accurately can help to resume operation of the network. The complexity of network technology has caused many difficulties to the fault diagnosis, it has important practical significance to carry out intelligent network fault diagnosis on the basis of the current research results.Many fine features of the artificial immune system open up new ways for the development of network fault diagnosis field, which get many researchers' attention. on the basis of extensive research about biological immune mechanisms, the case that mature detector can carry large amounts of information obtained in the evolution was found, the quality of mature detector group is critical to improve the diagnostic performance of the system. In addition, although the theory of memory cells is widely used in the immune theory research, memory cell populations is set to static, and there is no evaluation mechanism for memory cells. Therefore, the article starts to study the immune network theory applications in the field of fault diagnosis from the mature detector populations' quality and the memory detectors' memory characteristics.The subject study deeply of the biological immune characteristics and the DynamiCS' adaptability to the dynamic environment, based on negative selection, clonal selection and immune memory mechanism, referencing the operating mechanism of DynamiCS, an immune algorithm was proposed for network fault diagnosis, a network fault diagnosis model was designed based on immune theory. Clonal expansion strategy aimed at improving the quality of mature detector population. Memory classification strategy achieves memory detectors'dynamic update by the evaluation of memory detectors' memory characteristics. Mature and memory detector library in the fault diagnosis model complete evolution by using the clonal expansion strategy and memory classification strategy, the model can learn new emergence in the environment, generate new detector to detection fault, and improve the accuracy and efficiency of fault diagnosis effectively.