python参数函数二阶导数公式_python – 来自scipy.special的fadeeva函数的二阶导数

按照@Andras Deak的回答,您可以分析地找出高x扩展,然后使用一些简单的平滑在它和scipy函数之间进行插值.实际上有两个术语在高x扩展中取消,所以你必须要小心一点.

这是我得到的答案：

import numpy as np

import matplotlib.pyplot as plt

from scipy.special import wofz

def Z(x):

return wofz(x)

## first derivative of wofz (analytically)

def Zp(x):

return -2/1j/np.pi**0.5 - 2*x*Z(x)

def dawsn_expansion(x):

# Accurate to order x^-9, or, relative to the first term x^-8

# So when x > 100, this will be as accurate as you can get with

# double floating point precision.

y = 0.5 * x**-2

return 1/(2*x) * (1 + y * (1 + 3*y * (1 + 5*y * (1 + 7*y))))

def dawsn_expansion_drop_first(x):

y = 0.5 * x**-2

return 1/(2*x) * (0 + y * (1 + 3*y * (1 + 5*y * (1 + 7*y))))

def dawsn_expansion_drop_first_two(x):

y = 0.5 * x**-2