改进鸟群算法及其在发酵仿真建模中的研究  

An Improved Bird Swarm Algorithm and Its Application in Fermentation Simulation Modeling

作　　者：邵玉倩 刘登峰[1] 刘以安[1] SHAO Yu-qian;LIU Deng-feng;LIU Yi-an(School of Internet of Things,Jiangnan University,Jiangsu Wuxi 214122,China)

机构地区：[1]江南大学物联网工程学院

出　　处：《计算机仿真》2019年第11期220-227,共8页Computer Simulation

基　　金：国家自然科学基金青年项目(21706096);江苏省自然科学基金青年项目(BK20160162)

摘　　要：针对微生物发酵建模参数辨识过程中稳定性差、易陷入局部最优和模型预测精度低的问题,提出一种基于鸟群算法的发酵过程参数寻优算法-改进鸟群算法。通过采用非线性函数对原鸟群算法中的学习系数进行调整,并且当鸟类保持警戒行为并试图移动到种群中心时用莱维飞行公式替换鸟类位置更新公式,以及在寻优过程中当算法最优解保持不变时对最优解加入混沌扰动并用模拟退火算法再次寻优的三种方式对鸟群算法进行改进。仿真结果表明,改进鸟群算法在收敛速度、寻优精度和稳定性等方面的性能优于鸟群算法、遗传算法、粒子群算法等群智能算法。改进鸟群算法克服了原算法的不足之处,总体性能得到提高。To solve the problems of poor stability, local optimum, and low prediction accuracy in the parameters identification process of microbial fermentation, an Improved Bird Swarm Algorithm(IBSA) was proposed, which is a novel optimization algorithm for parameters of fermentation process based on the Bird Swarm Algorithm(BSA). The algorithm was improved in three ways. One was to adjust the learning coefficient in the original algorithm by using a nonlinear function, and the second was to replace the positional formula of the birds with the Levy flight formula when the birds remained vigilant and tried to move to the center of the population, and the third was to add chaotic perturbation to the optimal solution and optimize it again by the Simulated Annealing Algorithm(SAA) when the optimal solution of the algorithm remained unchanged during the optimization process. The experimental simulation results show that the IBSA is superior to the swarm intelligence algorithms such as BSA, Genetic Algorithm(GA) and Particle Swarm Optimization(PSO) in terms of convergence speed, optimization accuracy and stability. Therefore, the IBSA overcomes the shortcomings of the BSA and improves the overall performance.

关 键 词：发酵过程建模 鸟群算法 非线性调整 莱维飞行 混沌扰动 模拟退火 

分 类 号：TP301.6[自动化与计算机技术—计算机系统结构]