解方程组的单神经元自适应控制微粒群算法

Particle Swarm Optimization with the Neuron Adaptive Controller for Solving Equations

基于微粒群算法解决函数优化问题的优点,提出了使用微粒群算法求解方程组,并给出了求解方程组的通用模型。应用标准微粒群算法求解方程组容易陷入局部极值,导致方程组的解精度不高,并且算法具有较复杂的非线性特性。因此,将微粒群算法作为控制对象,引入单神经元控制器控制算法的惯性权重,将控制器具有的自学习、自适应能力和算法的全局优化特性相结合,用于方程组的求解。实验结果表明,该方法是有效可行的,适合于求解实际工程问题中的高非线性度方程组。

An effective optimization is proposed for solving equations, which combined particle swarm optimization with a single neuron intelligent controller. First, a common particle swarm optimization model is established. Then because the algorithm is easily trapped into local minimum and has a complicated nonlinear characteristic property, the dynamic inertia weight is changed, which uses a strong ability of self-studying and self-adaptating of a single neuron intelligent controller to improve the convergence rate and the convergence precision. The simulation results show that the method is effective and feasible, and give satisfactory solutions for solving systems of high nonlinear equations in factual engineering.