论文部分内容阅读
介绍了一种基于蒙特卡罗表述的空间缩减策略和局部边界线搜索的序列采样算法,该算法利用已有样本点的信息缩减原有设计空间,使得在缩减设计空间上生成的新样本点能够同时具有良好的空间填充特性和投影特性.与已有的序列采样算法的比较结果表明,该算法具有较高的采样效率和采样质量.采用此序列采样算法结合Kriging模型和遗传算法进行轮盘减质优化,优化结果减质10%.该序列采样算法为工程结构的优化提供了一条灵活有效的途径.
A Monte Carlo based spatial reduction strategy and local boundary line search sequence sampling algorithm is introduced. It uses the information of the existing sample points to reduce the original design space, so that the new sample points generated in the reduced design space can be And has good space filling characteristics and projection characteristics.Compared with the existing sequence sampling algorithm, this algorithm has high sampling efficiency and sampling quality.Using this sequence sampling algorithm combined with Kriging model and genetic algorithm for roulette subtraction The quality optimization and the optimization result are reduced by 10% .The sequence sampling algorithm provides a flexible and effective way to optimize the engineering structure.