基于高级遗传算法的船舶模型参数辨识

A Ships Parameter Identification Arithmetic Based on RGA

提出一种以二阶K-T方程为基础,并基于高级遗传算法的船舶参数辨识算法,能有效地保证了算法的收敛性和全局优化能力,避免陷入局部最优解。并引入噪声参数,再用基于多局部种群的遗传算法进行迭代计算,将残差白噪声化。最终将该方法与传统的广义最小二乘算法相比较,得到了较为理想的船舶参数。通过对采集的仿真数据和实际自动舵数据的辨识结果验证了该方法在实际工程应用中的可行性和有效性。

A ships parameter identification arithmetic based on two- order equations and adopted a refine genetic algorithm is proposed in this paper. This arithmetic is obviously improved on the element of the traditional genetic algorithm, can guarantee the astringency and globally optimized ability of the identification arithmetic and avoid sinking into a local optimized solution. Noise parameter is also introduced into the process, then iterative computations are done according to a genetic algorithm based on multi - local clusters, and the remained errors are transformed into white noises, finally the more ideal ships parameter is got compared with the traditional generalized least- squares algorithm. The possibility and the usefulness of this method were verified by the identification results of the conected simulation data and the actual automatic waterpilot data.