基于演化算法的带频率约束的桁架结构形状和尺寸优化  

Shape and size optimization of trusses with frequency constraints based on evolutionary algorithm

作　　者：薛运虎[1] 韦凌云[1] 赵玫[2] 郝仰梅[1] 唐天兵[3] 

机构地区：[1]北京邮电大学自动化学院,北京100876 [2]上海交通大学机械系统与振动国家重点实验室,上海200030 [3]广西大学计算机与电子信息工程学院,南宁530004

出　　处：《振动与冲击》2010年第12期13-17,共5页Journal of Vibration and Shock

基　　金：国家自然科学基金项目(50605010)

摘　　要：具有频率约束的桁架结构形状和尺寸优化设计是一个难度大的非线性动力优化问题。形状和尺寸变量的耦合通常导致收敛困难,而频率约束则使得动力灵敏度分析困难,传统的优化准则法和数学规划法难于求解。将单纯形算法、子空间变维技术、均匀变异有机融入郭涛算法,提出一种混合演化算法,避开繁琐的动力灵敏度分析,简单、有效地求解这类桁架形状和尺寸优化问题。典型的桁架算例验证了算法的有效性和可靠性。The shape and size optimization of trusses with frequency constraints is a highly nonlinear dynamic optimization problem.The coupling between two different types of design variables often leads to divergence,while multiple frequency constraints often cause difficulty in dynamic sensitivity analysis.The traditional mathematical programming and optimal criteria methods,which need complex dynamic sensitivity analysis and easily result in trapping into the local optima,are not sufficient to solve the problem.A modified evolutionary algorithm was proposed to solve the shape and size optimization of trusses simply and effectively.The proposed algorithm organically combines the Nelder Mead's simplex method,multiple-parent crossover and variant subspace technique.Several typical truss optimization examples were employed to demonstrate the validity,availability and reliability of the proposed algorithm for solving shape and size optimization of trusses with multiple frequency constraints.

关 键 词：桁架优化 演化算法 频率约束 结构优化 

分 类 号：TU311[建筑科学—结构工程]