基于混沌BBO算法的非线性模型参数整定  

Parameter estimating of nonlinear model based on chaotic BBO algorithm

作　　者：张滋雨 高岳林 武宏光[1] ZHANG Zi-yu;GAO Yue-lin;WU Hong-guang(School of Mathematics and Information Science,North Minzu University,Yinchuan 750021,China;Ningxia Key Laboratory of Intelligent Information and Data Processing,North Minzu University,Yinchuan 750021,China)

机构地区：[1]北方民族大学数学与信息科学学院,宁夏银川750021 [2]北方民族大学宁夏智能信息与大数据处理重点实验室,宁夏银川750021

出　　处：《计算机工程与设计》2023年第2期407-416,共10页Computer Engineering and Design

基　　金：国家自然科学基金项目(11961001、61561001);宁夏高等教育一流学科建设基金项目(NXYLXK2017B09);北方民族大学重大科研专项基金项目(ZDZX201901)。

摘　　要：为提升原始生物地理学优化算法(BBO)性能,提出基于动态迁移机制和混合变异算子的混沌生物地理学算法。采用Tent映射生成混沌初始化种群,提升种群遍历性;将反向学习机制和差分算子集成到原始迁移算子中,提升算法收敛速度;采用混合变异算子增强算法跳出局部最优解能力。将该算法应用于非线性Richards模型参数整定,预测谷氨酸菌体生长浓度。实验结果表明,该算法的预测结果比同类文献更优,较对比算法更适用于Richards模型参数整定。To improve the performance of biogeography-based optimization, a chaotic biogeography-based optimization based on dynamic migration mechanism and hybrid mutation mechanism was proposed. Tent mapping was used to generate chaotic initial population and improve ergodicity of population. The opposite-based learning and differential operator were integrated into the original migration operator to improve the convergence speed of the algorithm. The hybrid mutation operator was used to reduce the probability of the algorithm falling into the local optimal solution. The algorithm was applied to the parameter tuning of nonlinear Richards model to predict the growth concentration of glutamate. Numerical experimental results show that the prediction results of the proposed algorithm are superior to those of similar literature, and it is more suitable for Richards model parameter estimating than the comparison algorithm.

关 键 词：生物地理学优化算法(BBO) 动态迁移 混合变异 TENT映射 反向学习 Richards模型 参数整定 

分 类 号：TP181[自动化与计算机技术—控制理论与控制工程] TP306[自动化与计算机技术—控制科学与工程]