基于AMPSO-BP-GA的油船舱段结构优化  

Structure optimization of tanker cargo based on AMPSO-BP-GA

作　　者：王一镜 罗广恩[1] 刘家奇 刘俊成 WANG Yijing;LUO Guangen;LIU Jiaqi;LIU Juncheng(School of Naval Architecture and Ocean Engineering,Jiangsu University of Science and Technology,Zhenjiang 212100,China)

机构地区：[1]江苏科技大学船舶与海洋工程学院,镇江212100

出　　处：《江苏科技大学学报（自然科学版）》2023年第2期7-13,共7页Journal of Jiangsu University of Science and Technology:Natural Science Edition

基　　金：工信部高技术船舶基金资助项目;江苏省自然科学基金资助项目(BK20150468)。

摘　　要：由于船体结构的复杂性,传统优化方法容易出现陷入局部最优、求解速度偏慢等问题.基于自适应变异粒子群算法(adaptive mutation particle swarm optimization, AMPSO)、BP神经网络、遗传算法(genetic algorithm, GA),结合Isight/Nastran设计的正交试验,提出了AMPSO-BP-GA结构优化方法.以油船油货舱段结构的优化为算例,验证结构优化方法的有效性和可行性.结果表明:AMPSO算法相比于粒子群算法(particle swarm optimization, PSO)和GA算法具有更好的极值寻优能力;AMPSO-BP神经网络比PSO-BP和GA-BP预报精度更高;针对油船舱段结构优化,在相同的约束条件下,文中的AMPSO-BP-GA方法优化后重量较原设计减轻17.3%,优于GA-BP-GA优化方法的13.5%和PSO-BP-GA优化方法的13.4%.证明该方法具有可行性和有效性并具有推广性,可为船舶结构设计提供参考.Due to the complexity of the hull structure,traditional optimization methods are prone to problems such as falling into local optimum and slow solution speed.Based on adaptive mutation particle swarm optimization(AMPSO),BP neural network,genetic algorithm(GA),combined with orthogonal experiments designed by Isight/Nastran,the AMPSO-BP-GA structure optimization method is proposed.The results show that the AMPSO algorithm has better extreme value optimization capabilities than the PSO algorithm and the GA algorithm;the AMPSO-BP neural network has higher prediction accuracy than the PSO-BP and GA-BP;under the same constraint conditions,the proposed AMPSO-BP-GA method optimizes the structure so that the tank section is 17.3%lighter than the original design,which is better than the 13.5%weight reduction of the GA-BP-GA optimization method and the 13.4%that of the PSO-BP-GA optimization method.It is proved that the method is feasible,effective and generalizable,which can provide reference for other ship structure design.

关 键 词：舱段结构优化 BP神经网络 自适应变异粒子群算法 遗传算法 油船舱段 

分 类 号：U674.703.1[交通运输工程—船舶及航道工程]