介绍了求解抛物型波动方程的分步傅里叶变换(split step Fouriertransform,SSFT)算法计算过程,分析了算法的并行性,并基于西方快速傅里叶变换(fastest Fourier transform in the West,FFTW)函数库研究了2种分步傅里叶变换算法并行方案...介绍了求解抛物型波动方程的分步傅里叶变换(split step Fouriertransform,SSFT)算法计算过程,分析了算法的并行性,并基于西方快速傅里叶变换(fastest Fourier transform in the West,FFTW)函数库研究了2种分步傅里叶变换算法并行方案。所做测试结果表明,文中所提方案尤其是分布式模式方案,对于实现波动方程的快速求解是有效的,且所做工作对于以波动方程为基础的电波传播、电磁环境数据生成等问题的研究具有一定的指导意义。展开更多
In this article, two split-step finite difference methods for Schrodinger-KdV equations are formulated and investigated. The main features of our methods are based on:(i) The applications of split-step technique for S...In this article, two split-step finite difference methods for Schrodinger-KdV equations are formulated and investigated. The main features of our methods are based on:(i) The applications of split-step technique for Schrodingerlike equation in time.(ii) The utilizations of high-order finite difference method for KdV-like equation in spatial discretization.(iii) Our methods are of spectral-like accuracy in space and can be realized by fast Fourier transform efficiently. Numerical experiments are conducted to illustrate the efficiency and accuracy of our numerical methods.展开更多
文摘介绍了求解抛物型波动方程的分步傅里叶变换(split step Fouriertransform,SSFT)算法计算过程,分析了算法的并行性,并基于西方快速傅里叶变换(fastest Fourier transform in the West,FFTW)函数库研究了2种分步傅里叶变换算法并行方案。所做测试结果表明,文中所提方案尤其是分布式模式方案,对于实现波动方程的快速求解是有效的,且所做工作对于以波动方程为基础的电波传播、电磁环境数据生成等问题的研究具有一定的指导意义。
基金Supported by the National Natural Science Foundation of China under Grant No.11571181
文摘In this article, two split-step finite difference methods for Schrodinger-KdV equations are formulated and investigated. The main features of our methods are based on:(i) The applications of split-step technique for Schrodingerlike equation in time.(ii) The utilizations of high-order finite difference method for KdV-like equation in spatial discretization.(iii) Our methods are of spectral-like accuracy in space and can be realized by fast Fourier transform efficiently. Numerical experiments are conducted to illustrate the efficiency and accuracy of our numerical methods.