基于矩不确定分布鲁棒优化的船舶舱室布局设计

船舶舱室布局所涉的评价指标通常具有不确定性,诸如环境载荷等参数通常以不确定性的随机变量形式参与计算,且其概率分布的期望及方差等指标往往也是不确定的。本文采用矩不确定分布鲁棒优化方法,解决含矩不确定参数的船舶舱室布局min-max优化问题。将各舱室的全局位置参数、相对位置参数和质量分布参数加权整合为目标函数,建立全船范围内的舱室布局DRO-MU优化模型,采用拉格朗日对偶原理将其转为确定的半定规划问题以便求解。算例验证了本文方法的合理性与优越性,该方法针对性地用于解决含矩不确定参数的船舶舱室布局问题具有鲜明特点。

Ship Cabin Layout Design based on Distributional Robust Optimization under Moment Uncertainty

Evaluation indicators in the ship cabin layout problem usually present uncertainty. Parameters, such as environmental loads, usually participates calculation in the form of random variables. Meanwhile, the expectations and variances of their probability distribution are always uncertain. Distributional robust optimization under moment uncertainty (DRO-MU) is employed in this paper, in order to solve the ship cabin layout min-max optimization problem with moment uncertainty parameters. Global position parameter, relative position parameter and mass distribution parameter of each cabin is integrated as objective function, and DRO-MU model among the full range of ship is established. Furthermore, this optimization model is transformed into a deterministic semi-definite programming one by Lagrange duality principle. Calculation results show that the new method introduced here has rationality and superiority, and there are distinct characteristics for this method solving ship cabin layout problem with moment uncertainty parameters.