Research Of Coevolutionary Algorithm Based On Immune Agent

With the popularization of computer and network technology development,the computer network to people’s study, work and way of life has brought far-reaching influence. However, the development of network to provide moreconvenient conditions for the invaders.The invasion toward distributed,intelligent direction, the traditional network security technology and equipmenthas not fully protected against network attack. This paper studies the distributedintrusion detection system based on distributed intrusion detection system model,to do research and design, and the key problems--the intrusion detection modulethe evolution of key research done.Mobile Agent carrys out operation moved the software entities, it is usuallyin the way of autonomy the target host on running, it can real-time response tochanges in the environment, between Agent can cooperation, mutual defensedistributed attacks. Its independence, autonomy system are very good scalabilityand can be configured sex, it is the sexual response, can following the change ofenvironment and make corresponding response.In this paper a novel based on the model of equivalence distributed intrusiondetection system model. The model with the modern immunology theory andAgent technology is adopted the control center of many Agent collaborationstructure, composition of the system at the network’s Agent each node is flowing,real-time monitoring network status, and to identify their Agent each otherbehavior and the potential strategy according to take proper reaction, solved hassome distributed intrusion detection system between components dependent onbig, and intrusion detection system itself can not be automatically adapt to fixedstructure change of network intrusion and change, it is a kind of suitable forWAN assigned new distributed intrusion detection system model. Detailed modelthe overall design, including the function of each module, the parts of the internal structure, the work flow, the algorithm and the parts of mutual communicationand prevention, describes how the network data representation for the immunesystem of antigen, how to produce an immune to detect antigen antibody Agent,how to the communication between the Agent to distributed test, how to actinvasion response, etc.In order to solve the problem of common immune Agent evolution, thispaper puts forward a cooperative evolutionary mechanis m based on thedistributed collaborative Agent detection algorithm is proposed. This algorithmwas designed based on the distributed parallel mechanism of cooperativecoevolution of learning institutions and their agents coevolutionary algorithm,and makes the system agents in computerized control under the control of theinstitution, can parallel distribution coevolution local behavior rule library andglobal coordination behavior rule library. The system of inspection and testing ofthe local global cooperation has good online learning and adaptive characteristics,can be applied to practical resources limited embedded control system. Finally,using the traveler problems on the test, through the experimental results contrastand analysis, after verified the co-evolutionary algorithms in the distributedintrusion detection model of the detection function, detection effic iency isimproved and improve.