摘要
在一般免疫算法的基础上,提出了一种改进的多抗体集自适应免疫算法AIAMA。该算法把初始抗体集分成多个子抗体集和一个抗体池,每个子抗体集各自发展,体现了人体免疫系统的多样性。提出了抗体集活跃度的概念,对于具有不同活跃度的抗体集采用不同的增殖策略,既能保存优秀的抗体基因,又能跳出局部在广度上进行搜索。不同抗体集的活跃度可以随着环境动态改变,使算法具有自适应的特点。最后AIAMA分别在标准数据集和地下工程数据上进行了实验,结果表明了AIAMA算法的辨识精度更高、运算时间更短。
An improved adaptive immune algorithm ofmulti-antibodies （AIAMA） is inspired by the general immune algorithm. AIAMA divides the antibodies into several sub-antibody sets and one antibody pool, each sub-antibody set develops on its own which shows the diversity of the human immune system. The concept of ＇activity＇ is proposed that can decides the propagation strategy, in this way it can not only reserve the excellent antibody gene but also avoid local search, and the activity can adapt itself by sensing the change of the environment to make this algorithm adaptive. Finally the results of experiments based on the standard data set and underground engineering data set show that ALAMA has higher efficiency and quicker running time.
出处
《计算机工程与设计》
CSCD
北大核心
2010年第21期4661-4664,4681,共5页
Computer Engineering and Design
基金
国家自然科学基金项目(50778109)
上海市科技攻关计划基金项目(08511501702)
上海市重点学科建设基金项目(J50103)
关键词
免疫算法
多抗体集
活跃度
自适应
并行计算
immune algorithm
multi-antibody sets
activity
adaptive algorithm
parallel computing
