摘要
介绍了基于BP神经网络的全局性结构近似分析方法 ,解决了结构优化设计问题中变量的非线性映射问题。在此基础上 ,利用改进的遗传算法 ,对桁架结构在满足应力约束条件下进行结构最轻优化设计。利用Matlab的神经网络工具箱 ,编程求解了三杆桁架优化问题。
Optimal structural design method based on BP neural network and modified genetic algorithm were proposed in this paper.The high parallelism and non-linear mapping of BP neural network,an approach to the global structural approximation analysis was introduced.It can solve the mapping of design variables in structural optimization problems.Combining with an improved genetic algorithm,the truss structure is optimized to satisfy the full stress criteria.Under the condition of MATLAB 5.3,an example of truss structure has been solved by this method.
出处
《低温建筑技术》
2003年第1期34-36,共3页
Low Temperature Architecture Technology
关键词
BP神经网络
结构优化设计
满应力准则
衍架结构
Genetic algorithm
BP neural network
Structural optimization design
Full stress principle
