Research On Negative Select Immune Algorithm

Based on the study of selecting course of the biological immune system and Forrest negative selection algorithm, a new method is approached to improve the existing negative selection algorithm which has bad convergent effect in the space of string and high running costs. Then, the effectiveness of the method is analyzed. The arrangement of this paper is as follows:1. The problem of Forrest negative selection algorithm is analyzed in detail. The root cause of the problem is discussed with theory analysis and simulation experiments.2. As for the problem of Forrest negative selection algorithm running time and self scale become exponential relation, a detector library, from which the detector can be selected, is established based on the concept of pattern.. The detector library saved all the number of strings of unmatched self, so the space reduces a lot compared with all the string which unmatched self. And more running time can be saved when the size of self is larger, because the size of self becomes the line relations in running time.3. A new formula is employed to solve the problem. Detection failing ratiodeviates the expected value as the increase of the size of self, so the number of detectors adapt to the change of self scale.4. Based on the original algorithm, the improved method is analyzed and tested with simulation experiments. The result shows that the new algorithm has faster running speed of operation and lower detection failure rate compared with the original algorithm.5. The research work in this paper is summarized and the future work is approached.