Research Of The Intelligent Network Intrusion Detection Model

This paper has discussed the intelligent technology of intrusion detection, and has analyzed its current developed situation. According to the situation, this paper has proposed boldly the intelligent network intrusion detection model. In this model mature expert system becomes a main intrusion detection tool and GA-BP neural network only is an assisted one.To study this model, major work has been done as follows in this paper:First, this paper has analyzed three key technologies in the model, namely, expert system, neural network and genetic algorithm.Second, this paper uses genetic algorithm to optimize weights of BP neural network in order to achieve intrusion detection. There are three key technologies for the optimization: 1, chromosome spaces string is shined upon the weight coefficient; 2, the identification of fitness; 3, genetic operation.Third, generation rule device in this model also uses the genetic algorithm to generate rule. But this genetic algorithm is different from the genetic algorithm used to optimize BP nerve network. Their different point is that space of solution in this genetic algorithm isn't a simple number value and space of string, but a space of function. Therefore, the result from this genetic algorithm is not a simple constant string, but a variable string which can solve problem.