212 改良型免疫アルゴリズムの提案と評価

抄録

The Immune Algorithm (IA), which is optimum method based the human immune system, is similar to Genetic Algorithm (GA). The difference between the two methods is obtainable solution. .It's advantage for IA to gain several solutions at once. In addition, IA can expect to use the rocal rules which ware proposed for GA. On the other hands, IA may have the same demerit as GA, and be hard to deal with many parameters. In this paper, we suggest the improved Immune Algorithm. This method is improved by review of Immune System, and uses both secondary response and response contorol by endocrine system. It's hoped that new added functions affect searching ability. Additionally, the effect may be able to overleap complex parameters. This method is applied to some simple problems that search the extremum in the functions. As the results, it's suggestted the potential of overleaping parameter. The validity of this method is checked.