基于激素调节的量子免疫克隆函数优化算法  

Hormone adjustment based quantum-inspired immune clone algorithm for function optimization

作　　者：王毅[1,2] 孔晓琳[1] 牛奕龙[3] 齐敏[1,2] 樊养余[1,2] 

机构地区：[1]西北工业大学电子信息学院,陕西西安710072 [2]西北工业大学陕西省信息获取与处理重点实验室,陕西西安710072 [3]西北工业大学航海学院,陕西西安710072

出　　处：《系统工程与电子技术》2012年第9期1934-1939,共6页Systems Engineering and Electronics

基　　金：国家自然科学基金(60903127);西北工业大学基础研究基金(JC201108);西北工业大学"翱翔之星计划"基金资助课题

摘　　要：为了提高量子免疫克隆算法(quantum-inspired immune clone algorithm,QICA)对函数全局寻优的精确性和稳定性,引入了内分泌激素的调节规律,根据当前个体适应度值和上一代种群的平均适应度值重新设计克隆规模,按照种群多样性和Hill函数的上升规律对其进行自适应调整,使进化各代中优秀个体的克隆得到扩增,同时减少不良个体的规模,从而提出了一种基于内分泌激素调节的量子免疫克隆算法(hormone adjustmentbased QICA,HAQICA)。利用标准测试函数对算法进行了验证,50次随机独立实验结果表明,HAQICA算法的收敛速度与QICA算法相当,最优解的均值与方差等数据,证明了HAQICA算法在提高函数全局寻优性能上的有效性。Based on the adjustment regulation of endocrine hormone, a novel algorithm, called the hormone adjustment based quantum-inspired immune clone algorithm （HAQICA）, is proposed to improve the accuracy and stability of quantum-inspired immune clone algorithm （QICA） on global optimization. In HAQICA, the clone size is calculated according to the individual fitness of the current generation and the average fitness of the previous generation, and is adjusted adaptively in terms of the population diversity and the rise law of Hill func- tion that is the basic model of endocrine networks. HAQICA also increases the clone number of better individ- uals and decreases the clone number of worse individuals. Standard test functions are used to verify the algo- rithm, and the results of 50 random independent experiments show that the convergence speed of HAQICA is comparative with that of QICA and HAQICA is more efficient in global optimization according to the mean and variance values of optimal solutions.

关 键 词：量子免疫克隆算法 激素调节规律 克隆规模 函数优化 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]