求解非线性方程组的迭代神经网络算法

求解非线性方程组是工程研究中的基本问题,普通的求解算法均具有一定的缺点,通用性不强。神经网络能以任意精度逼近非线性函数,利用它逼近非线性方程组的函数的反函数,提出了通用性较强的数值求解方法。首先,给出了不需迭代的简单神经网络算法;然后,针对给定求解区域偏大和不准确的问题,提出了缩小与改变求解区域的迭代神经网络算法。这两种算法均进行了实例求解,结果表明,两种算法格式简单,求解时间短,精度高,具有较高的应用价值,在理论研究和工程实践中具有较大应用前景。最后分析了算法的优点和改进方向。

Iterative algorithms of neural network for nonlinear equation groups

Solving nonlinear equation groups is a principal problem in engineering study.Common solving algorithms all have their disadvantages.Because neural network can precisely approach any function,it is used to approach the inverse function of function of nonlinear equation groups,this dissertation giving numerical algorithm for the groups.First,a simple algorithm of neural network is given.In order to solve problems of big or wrong given area of roots,a iterative algorithm is advanced by reducing or changing th...