改进生物地理学算法辨识Hammerstein模型.pdfVIP

摇 摇 第31卷摇 第5期 计摇 算摇 机摇 仿摇 真 2014年5月摇 摇 文章编号:1006-9348(2014)05-0342-04 改进生物地理学算法辨识Hammerstein模型 罗摇 丹,张宏立 (新疆大学电气工程学院,新疆 乌鲁木齐830047) 摘要:为了研究非线性控制的辨识优化问题,提出辨识非线性模型的新方法。 针对传统辨识方法对非线性模型中非线性部 分辨识精度低、效果差等缺点,将生物地理学算法引入到非线性辨识中,并结合Hammerstein模型,提出了一种基于生物地理 学算法的Hammerstein模型识别方法。 为了进一步增强生物地理学算法的辨识精度,采用了带有加权因子的迁移和基于柯 西变异的改进生物地理学算法对非线性系统进行辨识。 通过数值仿真,并与传统的最小二乘法和基本生物地理学算法进行 比较,结果表明利用改进算法能够提高辨识精度,验证了改进算法的有效性和可行性。 关键词:生物地理学优化算法;系统辨识;海默斯坦模型 中图分类号:TP202+.7摇 摇 文献标识码:A Identification of Hammerstein Model Based on Improved Biogeography-Based Optimization Algorithm LUO Dan,ZHANG Hong-li (College of Electrical Engineering,Xinjiang University,Urumqi Xinjiang830047,China) ABSTRACT:A new parameteridentificationmethodfornonlinearmodelwasproposed. Inordertoovercomethedis鄄 advantage that the recognition accuracy for the nonlinear part of nonlinear model is poor in traditional identification methods,the Biogeography-Based Optimization (BBO) was introduced to nonlinear system identification and com鄄 bined Hammerstein model,andthe Hammersteinmodel identificationmethodbasedon BBOalgorithmwasproposed. In order to enhance the identificationperformanceof the BBOalgorithm,the BBOwasimprovedbasedonpartial mi鄄 gration with weight factor and the Cauchy mutation was used to identify nonlinear system. According to a numerical simulation and comparingwiththetraditional least squaremethod andthebasic BBOalgorithm,theresults show that the algorithm can improved the identification accuracy andobtain agoodrecognitionresults,whichverifiesthe effec鄄 tiveness andfeasibility of the algorithm. KEYWORDS:Biogeography-based optimization