船型论证中的随机仿真与优化

Optimization of Stochastic Economic Demonstration of Merchant Ships by SAM

随机性是运输系统的主要特性之一。但是,过去由于随机系统的优化手段不够成熟,计算量相当大,因此,在实际优化过程中不得不作了许多简化与假设。这样对最后所得到的优化结果的准确性与可靠性会产生一定的不利影响。本文结合受随机干扰影响的船型论证工作,介绍了一种连续型随机系统优化算法——随机逼近法SAM,它比传统的建立在均值处理基础上的仿真优化算法——直接搜索法DSM,无论在运算速度还是计算准确性等方面均有显著的优越性,也比通常只能逐个分析处理备选方案而不能得出最优解的不确定性分析与风险分析方法更方便,有效和全面。

Transpertation system is a kind of stochastic system. But it is difficult and time-consuming to optimize stochastic models with Sensitivity Analysis, probability Analysis or the Direct Search Methods(DSM). Hence, transportation system is usually handled as a kind of deterministic system. This paper introduces the Stochastic Approximation Methods(SAM)in the economic demonstatrion of merchant ships under certain conditions, which obtains satisfactory optimization results.