scipy.special.airye#

scipy.special.airye(z, out=None) = <wrapped_ufunc 'airye'>#

Exponentially scaled Airy functions and their derivatives.

Scaling:

eAi  = Ai  * exp(2.0/3.0*z*sqrt(z))
eAip = Aip * exp(2.0/3.0*z*sqrt(z))
eBi  = Bi  * exp(-abs(2.0/3.0*(z*sqrt(z)).real))
eBip = Bip * exp(-abs(2.0/3.0*(z*sqrt(z)).real))

Parameters:

zarray_like: Real or complex argument.
outtuple of ndarray, optional: Optional output arrays for the function values

Returns:

eAi, eAip, eBi, eBip4-tuple of scalar or ndarray: Exponentially scaled Airy functions eAi and eBi, and their derivatives eAip and eBip

See also

airy

Notes

Wrapper for the AMOS [1] routines zairy and zbiry.

Array API Standard Support

airye has support for Python Array API Standard compatible backends in addition to NumPy. The following combinations of backend and device (or other capability) are supported.

Library	CPU	GPU
NumPy	✅	n/a
CuPy	n/a	⛔
PyTorch	✅	⛔
JAX	✅	⛔
Dask	⛔	n/a

For the NumPy backend, this function supports all NumPy ufunc keyword arguments. Other backends may support out, but typically none of the other ufunc kwargs. out is typically supported for CuPy and PyTorch, but not currently in cases where SciPy relies on a generic Array API implementation or, for PyTorch on CPU, falls back to the NumPy backend. out is never supported for JAX because JAX arrays are immutable. airye does not currently support out for the PyTorch backend.

See Support for the array API standard for more information.

References

[1]

Donald E. Amos, “AMOS, A Portable Package for Bessel Functions of a Complex Argument and Nonnegative Order”, http://netlib.org/amos/

Examples

We can compute exponentially scaled Airy functions and their derivatives:

>>> import numpy as np
>>> from scipy.special import airye
>>> import matplotlib.pyplot as plt
>>> z = np.linspace(0, 50, 500)
>>> eAi, eAip, eBi, eBip = airye(z)
>>> f, ax = plt.subplots(2, 1, sharex=True)
>>> for ind, data in enumerate([[eAi, eAip, ["eAi", "eAip"]],
...                             [eBi, eBip, ["eBi", "eBip"]]]):
...     ax[ind].plot(z, data[0], "-r", z, data[1], "-b")
...     ax[ind].legend(data[2])
...     ax[ind].grid(True)
>>> plt.show()

../../_images/scipy-special-airye-1_00_00.png

We can compute these using usual non-scaled Airy functions by:

>>> from scipy.special import airy
>>> Ai, Aip, Bi, Bip = airy(z)
>>> np.allclose(eAi, Ai * np.exp(2.0 / 3.0 * z * np.sqrt(z)))
True
>>> np.allclose(eAip, Aip * np.exp(2.0 / 3.0 * z * np.sqrt(z)))
True
>>> np.allclose(eBi, Bi * np.exp(-abs(np.real(2.0 / 3.0 * z * np.sqrt(z)))))
True
>>> np.allclose(eBip, Bip * np.exp(-abs(np.real(2.0 / 3.0 * z * np.sqrt(z)))))
True

Comparing non-scaled and exponentially scaled ones, the usual non-scaled function quickly underflows for large values, whereas the exponentially scaled function does not.

>>> airy(200)
(0.0, 0.0, nan, nan)
>>> airye(200)
(0.07501041684381093, -1.0609012305109042, 0.15003188417418148, 2.1215836725571093)