| 引用本文: | 张波涛,刘士荣,吕强.采用生物信息克隆的免疫算法[J].控制理论与应用,2010,27(6):799~803.[点击复制] |
| ZHANG Bo-tao,LIU Shi-rong,LV Qiang.Immune algorithm with biologic information clone[J].Control Theory and Technology,2010,27(6):799~803.[点击复制] |
|
| 采用生物信息克隆的免疫算法 |
| Immune algorithm with biologic information clone |
| 摘要点击 2002 全文点击 1091 投稿时间：2009-05-13 修订日期：2009-08-10 |
| 查看全文 查看/发表评论 下载PDF阅读器 |
| DOI编号 10.7641/j.issn.1000-8152.2010.6.CCTA090609 |
| 2010,27(6):799-803 |
| 中文关键词 智能计算 人工免疫系统 克隆选择 |
| 英文关键词 intelligent computing artificial immune system clonal selection |
| 基金项目 国家自然科学基金资助项目(60675043); 浙江省科技计划资助项目(2007C21051); 浙江省自然科学基金资助项目(Y1090426). |
|
| 中文摘要 |
| 受克隆选择过程生物学原理的启发, 提出了一种采用生物信息克隆的免疫算法. 抗体克隆依赖于一个动态平衡的网络, 并与遗传因素相关. 为了解决传统克隆过程中信息不能充分利用的问题, 该进化算法将环境信息、抗体历史信息以及抗体遗传特征积累的影响引入人工免疫系统, 用这多种信息作为先验知识为克隆过程提供决策支持, 引导抗体系统的更新. 同时采用实数与二进制混合编码方式增加种群多样性, 提高收敛速度, 然后分析了该算法的收敛性. 仿真实验结果表明, 该克隆策略能较大的提高免疫克隆算法的优化能力; 与几种高级免疫克隆算法和进化算法相比, 该算法寻优精度高, 收敛速度快, 能有效的克服早熟现象, 并具有很好的高维优化能力. |
| 英文摘要 |
| The clonal selection process depending on a homeostasis network is closely linked with genetic factors. Inspired by this immunology principle, an immune algorithm with biologic information clone(IABIC) is proposed. To deal with the missing information in traditional artificial clonal selection process, this evolutionary algorithm introduces the environment information, history information of antibody and hereditary characteristics into the artificial immune system, which serves as the prior knowledge to provide decision support for clonal selection process and conduct the rebirth of antibody system. Meanwhile, hybrid coding was adopted to improve the diversity of the population and the convergence speed, and then, the global convergence of this algorithm was analyzed. Simulation results indicate that this clonal strategy makes significant improvement in optimizing the ability of immune clonal algorithm. Compared with other immune clonal optimization algorithms and evolution algorithms, it has higher precision, faster convergence, remarkable convergence character for higher dimensional optimization, and the ability for avoiding premature convergence. |
|
|
|
|
|