基于Hopfield神经网络的结构优化算法  

Structural optimization algorithm based on Hopfield neural network

作　　者：李宇洁[1] 李章政[1] 

机构地区：[1]四川大学建筑与环境学院,四川成都610065

出　　处：《四川建筑科学研究》2013年第1期27-30,共4页Sichuan Building Science

摘　　要：利用Hopfield神经网络与增广拉格朗日乘子法相结合来求解非线性约束的结构优化问题。针对神经网络模型容易陷入局部极小解的缺点,网络引入了模拟退火算法,提高了网络全局寻优能力,得到了较好的优化结果,本文详述了算法的步骤,编制了该算法的计算程序,最后通过两个实例证明该算法的可行性和有效性。Use Hopfield neural network and augmented Lagrange multiplier method to solve nonlinear constraint with the structure optimization problems. According to the neural network model easily into the local minimum solution shortcomings, network into the simulated annealing algorithm,improve the network global optimization ability and obtain the better the optimization results ,this paper described the algorithm of steps, the algorithm of the calculation procedure, finally through two examples show the feasibility and effectiveness of the algorithm.

关 键 词：HOPFIELD神经网络 结构优化 增广拉格朗日乘子法 模拟退火 

分 类 号：TU37[建筑科学—结构工程]