摘要
提出了一种自适应免疫算法,其很好地解决了原始免疫算法中收敛精度低和寻优速度慢的缺点。通过对比分析标准测试函数的计算结果,自适应免疫算法的优良性得到充分地证明。然后,免疫算法被用于优化BP神经网络的结构和参数。结果表明,不但网络结构得到较好地控制,而且泛化性能也有较大地提高。最后,算法在优化神经网络上的有效性也在焦化装置精馏塔汽油干点软测量建模中得到很好地证实。
In this paper, a self-adaptive immune algorithm （SAIA） problems of slow convergence speed and low calculation precision of the was presented which well addressed the original immune algorithm. By analyzing the results of bench mark functions, the excellent performance of SAIA was proved. Then, the SAIA was applied to optimize the structure and parameters of BPNN. And the availability of algorithm optimizing neural network was proved by applying BPNN in gas concentration soft sensor modeling of coke sets.
出处
《广州化工》
CAS
2009年第1期17-19,34,共4页
GuangZhou Chemical Industry
关键词
免疫算法
BP神经网络
软测量
immune algorithm
BP neural network
soft sensor
