基于分阶段搜索连续蚁群算法的船舶纵向运动参数辨识  

Parameter identification of ship longitudinal motions using continuous ant colony algorithm with period searching

作　　者：戴运桃[1,2] 赵希人[2] 刘利强[2] 

机构地区：[1]哈尔滨工程大学理学院,哈尔滨150001 [2]哈尔滨工程大学自动化学院,哈尔滨150001

出　　处：《船舶力学》2010年第8期872-878,共7页Journal of Ship Mechanics

基　　金：国防科学技术工业委员会基础研究基金资助项目(41314020201)

摘　　要：提出了一种分阶段搜索的连续蚁群算法,并成功应用于求解船舶纵向运动参数辨识问题。首先将船舶纵向运动的参数辨识问题转化为参数空间非线性优化问题,然后在优化问题求解过程中,依据待辨识参数对待优化问题影响的大小,将所有参数进行动态分组,依据影响由大到小的顺序,利用连续蚁群算法依次对各组参数进行寻优,确定各组参数的范围,最后对所有参数进行小范围精细搜索,从而使算法最终收敛到最优解。求解结果表明,该算法能够快速地辨识出满足精度要求的船舶纵向运动水动力参数,验证了算法的有效性。A continuous ant colony algorithm with period searching（PSCACA） is introduced.It was suc-cessfully applied in solving the problem of hydrodynamic parameter identification of ship longitudinal mo-tions.Firstly,the problems of parameter identification of ship longitudinal motions are converted to nonlin-ear optimization problems in parameter space,and then,in the solving of the optimization problem,the pa-rameters are dynamic grouped in the light of the importance of each parameter respectively.Furthermore,the ranges of the parameter are fixed and the search is practiced with high accuracy in these ranges.The result of hydrodynamic parameter identification of ship shows that the algorithm can recognize the hydro-dynamic parameter of ship longitudinal motions in a satisfied precision.And the algorithm is proved to be effective.

关 键 词：参数辨识 分阶段搜索 连续蚁群 纵向运动 

分 类 号：TP301.6[自动化与计算机技术—计算机系统结构]