Intelligent optimization algorithms to VDA of models with on/off parameterizations

A useful approach to sensitivity and predictability studies in geophysical fluid dynamics: conditional non-linear optimal perturbation

“…This type of method was thus also employed to compute the CNOP. For example, the genetic algorithm (GA, [ 66 ]) and particle swarm optimization (PSO, [ 67 ]) have been used to obtain the CNOP for models with different degrees of freedom, such as the theoretical Lorenz model and the 2D Ikeda model [ 41 , 68 , 69 ]. The larger dimensional (240-dimension) optimization problem [ 70–72 ] was solved to determine the CNOP related to parameters using the differential evolution (DE, [ 73 ]) algorithm.…”

Quantitative Precipitation Forecast of a Tropical Cyclone through Optimal Parameter Estimation in a Convective Parameterization

“…The GAs are based on natural genetic and natural selection mechanism, and some fundamental ideas are borrowed from Genetics in order to artificially construct an optimization procedure (Holland 1992). GAs have been successfully applied to a range of atmospheric problems and have characteristics of easy implementation and capability of achieving global optimal solution (Fang et al 2009;Kishtawal et al 2003;Singh et al 2005a, b).…”

一种有效求解含有不连续“开关”过程的变分资料同化的遗传算法

“…Fang 等 [34] 将粒子群最优化(PSO)和 GA-ED 两种 优化算法用以修改的 Lorenz 模型的变分资料同化试 验, 结果显示 PSO 和 GA-ED 均有较强的寻优能力, 且 PSO 与 GA-ED 相比较, 在计算耗时及对观测误差 的鲁棒性方面更有优势. 但标准的 PSO 方法的收敛 速度仅在对低维向量优化时快(如对 Lorenz 模型的变 分资料同化仅仅是对三维向量寻优), 而随着优化问 题的维数的增加, 其收敛速度会快速下降.…”