基于混合优化策略的多学科协同优化及其应用  

Application of a new collaborative optimization based on mixed strategy

作　　者：许辉[1] 周奇[1] 张立丽[2] 

机构地区：[1]中国舰船研究设计中心,湖北武汉430064 [2]江苏自动化研究所,江苏连云港222006

出　　处：《舰船科学技术》2014年第11期23-28,共6页Ship Science and Technology

基　　金：国防基础科研重点资助项目(A0820110001)

摘　　要：多学科设计优化为复杂系统工程设计提供了新的思路并成为优化设计领域的研究热点。针对传统多学科协同优化算法常常出现无法收敛或收敛陷入局部最优的问题,提出基于模拟退火算法和序列二次算法的混合协同优化(collaborate optimization based on simulated annealing and sequential quadratic programming,SA-SQP-CO),SA-SQP-CO应用模拟退火算法和序列二次算法的混合优化策略取代传统系统级基于梯度的求解方法,同时引入动态松弛的思想,在二阶段寻优过程中采用动态松弛量代替系统级一致性等式约束加强学科一致性、提高系统级收敛效率。以经典齿轮减速箱测试算例,通过与传统多学科协同优化算法比较,验证了该方法在优化结果可靠性、稳定性等方面有优势。最后,应用SA-SQP-CO算法求解抛物线型载荷下纤维加强悬臂梁轻量化设计问题以体现其工程实用性。Multidiscipline design optimization provides a promising methodlogy for largr-scale system design and becomes an active field of optimization research. Facing the shortcomings of traditional collaborative optimization, such as time-consuming, being sensitive to the initial points ,not converging . A new collaborative optimization based on simulated annealing and sequential quadratic programming is presented. Firstly ,replacing the gradient-based method by the hybrid optimization strategy of the simulated annealing algorithm and sequential quadratic programming in system-level. Secondly, dynamic relaxation factor used in the optimization process to improve the convergence rate. Comparing with the traditional collaborative optimization methods via gearbox problem, the better convergence, stability and reliability of the presented collaborative optimization are demonstrated. Finaly,the new collaborative optimization is used to solve a continuous fiber-reinforced composite cantilever beam subject to a parabolic distributed load and a satisfied optimization result is also achieved.

关 键 词：协同优化 混合策略 模拟退火 动态松弛量 悬臂梁 

分 类 号：U662[交通运输工程—船舶及航道工程] O224[交通运输工程—船舶与海洋工程]