NAG Library Function Document

nag_complex_erfc (s15ddc)

. Following advice in Gautschi (1970) and van der Laan and Temme (1984), the code in Gautschi (1969) has been adapted to work in various precisions up to

18

decimal places. The real part of

w (z)

is sometimes known as the Voigt function.

4 References

Gautschi W (1969) Algorithm 363: Complex error function Comm. ACM 12 635

Gautschi W (1970) Efficient computation of the complex error function SIAM J. Numer. Anal. 7 187–198

van der Laan C G and Temme N M (1984) Calculation of special functions: the gamma function, the exponential integrals and error-like functions CWI Tract 10 Centre for Mathematics and Computer Science, Amsterdam

5 Arguments

1: z – ComplexInput: On entry: the argument $z$ of the function.
2: fail – NagError *Input/Output: The NAG error argument (see Section 3.6 in the Essential Introduction).

6 Error Indicators and Warnings

NE_INTERNAL_ERROR: An internal error has occurred in this function. Check the function call and any array sizes. If the call is correct then please contact NAG for assistance.
NE_RESULT_HALF_PRECISION: Result has less than half precision when entered with argument $z = (⟨value⟩, ⟨value⟩)$ .
NE_RESULT_IMAGINARY_OVERFLOW: Imaginary part of result overflows when entered with argument $z = (⟨value⟩, ⟨value⟩)$ .
NE_RESULT_NO_PRECISION: Result has no precision when entered with argument $z = (⟨value⟩, ⟨value⟩)$ .
NE_RESULT_OVERFLOW: Both real and imaginary parts of result overflow when entered with argument $z = (⟨value⟩, ⟨value⟩)$ .
NE_RESULT_REAL_OVERFLOW: Real part of result overflows when entered with argument $z = (⟨value⟩, ⟨value⟩)$ .

7 Accuracy

The accuracy of the returned result depends on the argument

z

. If

z

lies in the first or second quadrant of the complex plane (i.e.,

Im (z)

is greater than or equal to zero), the result should be accurate almost to machine precision, except that there is a limit of about

18

decimal places on the achievable accuracy because constants in the function are given to this precision. With such arguments, fail can only return as

fail . code =

NE_NOERROR.

If however

Im (z)

is less than zero, accuracy may be lost in two ways; firstly, in the evaluation of

e^{- z^{2}}

, if

Im (- z^{2})

is large, in which case a warning will be issued through

fail . code =

NE_RESULT_HALF_PRECISION or NE_RESULT_NO_PRECISION; and secondly, near the zeros of the required function, where precision is lost due to cancellation, in which case no warning is given – the result has absolute accuracy rather than relative accuracy. Note also that in this half-plane, one or both parts of the result may overflow – this is signalled through

fail . code =

NE_RESULT_IMAGINARY_OVERFLOW, NE_RESULT_OVERFLOW or NE_RESULT_REAL_OVERFLOW.

8 Parallelism and Performance

Not applicable.

9 Further Comments

The time taken for a call of nag_complex_erfc (s15ddc) depends on the argument

z

, the time increasing as

|z| \to 0.0

nag_complex_erfc (s15ddc) may be used to compute values of

erfc z

and

\erf z

for Complex

z

by the relations

erfc z = e^{- z^{2}} w (i z)

\erf z = 1 - erfc z

. (For double arguments, nag_erfc (s15adc) and nag_erf (s15aec) should be used.)

10 Example

This example reads values of the argument

z

from a file, evaluates the function at each value of

z

and prints the results.

NAG Library Function Documentnag_complex_erfc (s15ddc)

+− Contents

1 Purpose

2 Specification

3 Description