CAJ | 학술논문

提出用径向基神经网络求解第二类Fredholm方程的方法.首先使用径向基神经网络逼近积分方程中的未知函数,然后将求解第二类积分方程转化为一个优化问题.粒子群优化算法具有不易陷入局部极小、易实现和调整参数较少的优点,从而利用粒子群优化算法的求解该优化问题.数值实验表明所提方法是可行的.
제출용경향기신경망락구해제이류Fredholm방정적방법.수선사용경향기신경망락핍근적분방정중적미지함수,연후장구해제이류적분방정전화위일개우화문제.입자군우화산법구유불역함입국부겁소、역실현화조정삼수교소적우점,종이이용입자군우화산법적구해해우화문제.수치실험표명소제방법시가행적.
In this paper, the method based on radial basis function neural networks (RBFNN) is presented to solve the second kind linear Fredholm integral equations. Firstly, RBFNN is applied to approximate the un-known function in integral equations, and then solving the second kind Fredholm can be converted to solve an optimization problem. The advantages of Particle swarm optimization (PSO) are that PSO possesses the capa-bility to escape from local optima,and it is easy to be implemented and has fewer parameters to be adjusted.The solved optimization problem can be found solutions using PSO. Numerical experiment shows the proposed methods yield better results, and so our method is feasible and efficient.