摘要
提出一种以费用为约束条件的连续状态并串联系统冗余度优化设计模型。用传统的优化算法很难求得该模型目标函数的显性表达式,文中提出用BP(back propagation)神经网络对目标函数进行逼近,得到的近似目标函数是标准sigmoid函数的线性组合。仿真结果表明,该方法可以精确地求得系统的最优冗余度。
An optimal redundancy design model of continuous-state parallel-series systems with the constraint of cost was presented. It was very hard to obtain an explicit expression of the objective function for the model by using conventional optimization algorithms. A BP（back propagation） neural network approach was proposed to approximate the objective function. The obtained approximate objective function is a linear combination of standard sigmoid functions. The simulation result shows that the optimal redundancy of a system can be obtained accurately by this method.
出处
《机械强度》
EI
CAS
CSCD
北大核心
2007年第2期247-250,共4页
Journal of Mechanical Strength
关键词
可靠性
BP神经网络
连续状态
冗余度优化设计
sigmold函数
逼近
Reliability
BP（back propagation） neural network
Continuous-state
Optimal redundancy design
Sigmoid function
Approximate
