融合克隆选择的AEA算法及其在参数估计中的应用  

The AEA combined with clonal selection algorithmand its application on parameter estimation

作　　者：桑志祥[1] 李绍军[1] 董跃华[1] 付煜[1] 

机构地区：[1]华东理工大学自动化系,上海200237

出　　处：《计算机与应用化学》2011年第12期1527-1530,共4页Computers and Applied Chemistry

基　　金：国家自然科学基金资助项目(20976048.21176072)

摘　　要：为了提高AEA算法的寻优性能,提出了结合克隆选择算法改进AEA算法计算过程中每代种群的生产方式。改进后的AEA算法(AEA-C)的克隆选择部分所加入的噪声是随着实际进化不同阶段而合理地变化。在进化的前期,由于添加噪声较大有利于算法的全局搜索,随着迭代次数增加,噪声逐渐减小,使得算法加强了局部搜索,算法能跳出局部最优,避免早熟现象的发生。在10个典型测试函数上进行了试验,结果表明AEA-C的寻优性能有了很大的提高,不仅获得的解的质量好,而且算法的运算速度和稳定性都得到了提高。最后将AEA-C算法应用于发酵动力学模型参数的估计,效果明显。Considering the unreasonable method in AEA that generates candidate solutions at each iteration, and clonal selection algorithm can improve that way of generating population, an improved AEA algorithm （AEA-C） which is fused AEA with clonal selection algorithm is proposed in this paper. The noise which is added to the clonal selection part of proposed algorithm reasonably changes with different stages of practical evolution. In the early evolution, the noise is large enough to strengthen the global search. With the increase in the number of iteration, the noise decreases, making algorithm enhance global search to help it can escape local optima, and avoid premature phenomenon. The performance of the proposed new algorithm is studied with the use of 10 benchmark functions and shows that the AEA-C clearly outperforms the other algorithms for almost all the 10 benchmark functions. Not only is good solution obtained, but also computing speed and stability of the algorithm are improved. Finally, it is applied to the parameter estimation of fermentation dynamics models, and the satisfactory results are obtained.

关 键 词：ALOPEX AEA 克隆选择算法 优化 参数估计 

分 类 号：TQ015.9[化学工程] TP391.9[自动化与计算机技术—计算机应用技术]