基于Hopfield神经网络和KS函数的优化设计研究

介绍了Hopfield神经网络优化的原理。将神经网络的求解转化为非线性微分方程组的初值问题，并利用MATLAB提供的微分方程组求解器进行计算。对于多约束优化问题，KS函数的包络凝聚特性可以简化约束条件，其光滑可微特性又方便了问题的求解。将其运用于非线性约束规划和某型飞机总体参数优化问题，算例表明此方法是有效的。

Optimization Design Based on Hopfield Neural Network & KS-Function

The principle of Hopfield neural network for optimization is presented. The numerical implementation of the network is transformed into the initial value problem of non-linear differential equations, and the ODE solver of MATLAB can be used. For multi-constraint optimization problems, KS function can agglomerate these constraints, the smooth and differentiable feature of which is favorable for the solution. The results from the application of the method to a non-liner constrained optimization and a parameter optimization for the conceptual design of an aircraft indicate that the method is effective.