The Improvement Of Genetic Algorithm And It's Application In IDS

Genetic Algorithm(GA) uses colony search techniques to generate a newpopulation which includes the best or an approximate solution for a special givenproblem by applying "Selection"," Crossover" and "Mutation" to the currentpopulation. The GA is widely used in self-adaptive control, combinationoptimization, pattern recognition, machine learning, artificial life, management anddecision making, etc. The main reason for this is it is easy to be implemented, robustand good-performance.Intrusion Detection System(IDS) is an important part of network securitymanagement. It judges whether a network connection is legality or not via the rulesand regulations. In this study, the GA was used to the IDS to form new rulesautomatically based on prior defined historical network connection data. Unlikeother implementations, the solution proposed in this work considers both timing andspatial information of network connections, which these two factors will be classifiedinto the rules in IDS. Such kind of the regulations and rules will be significant toidentify any complex abnormal behaviors.This work is focused on the TCP/IP network protocols. In the tested problem,proposed technique can give multiple optimized solutions. Based on the tested data,it worked well.