Nag_BoundInterval boundinf, double bound, double epsabs, double epsrel, Integer max_num_subint, double *result, double *abserr, Nag_QuadProgress *qp, Nag_User *comm, NagError *fail)

The function may be called by the names: d01smc, nag_quad_dim1_quad_inf_1 or nag_1d_quad_inf_1.

3 Description

d01smc is based on the QUADPACK routine QAGI (Piessens et al. (1983)). The entire infinite integration range is first transformed to

[0, 1]

using one of the identities

\int_{- \infty}^{a} f (x) d x = \int_{0}^{1} f (a - \frac{1 - t}{t}) \frac{1}{t^{2}} d t

\int_{a}^{\infty} f (x) d x = \int_{0}^{1} f (a + \frac{1 - t}{t}) \frac{1}{t^{2}} d t

\int_{- \infty}^{\infty} f (x) d x = \int_{0}^{\infty} (f (x) + f (- x)) d x = \int_{0}^{1} [f (\frac{1 - t}{t}) + f (\frac{- 1 + t}{t})] \frac{1}{t^{2}} d t

where

a

represents a finite integration limit. An adaptive procedure, based on the Gauss 7-point and Kronrod 15-point rules, is then employed on the transformed integral. The algorithm, described by de Doncker (1978), incorporates a global acceptance criterion (as defined by Malcolm and Simpson (1976)) together with the

ε

-algorithm (Wynn (1956)) to perform extrapolation. The local error estimation is described by Piessens et al. (1983).

4 References

de Doncker E (1978) An adaptive extrapolation algorithm for automatic integration ACM SIGNUM Newsl. 13(2) 12–18

Malcolm M A and Simpson R B (1976) Local versus global strategies for adaptive quadrature ACM Trans. Math. Software 1 129–146

Piessens R, de Doncker–Kapenga E, Überhuber C and Kahaner D (1983) QUADPACK, A Subroutine Package for Automatic Integration Springer–Verlag

Wynn P (1956) On a device for computing the

e_{m} (S_{n})

transformation Math. Tables Aids Comput. 10 91–96

5 Arguments

1: $f$ – function, supplied by the user External Function

f must return the value of the integrand

f

at a given point.

The specification of f is:

double

f (double x, Nag_User *comm)

1: $x$ – double Input

On entry: the point at which the integrand

f

must be evaluated.

2: $comm$ – Nag_User *

Pointer to a structure of type Nag_User with the following member:

p – Pointer: On entry/exit: the pointer $comm \to p$ should be cast to the required type, e.g., struct user *s = (struct user *)comm → p, to obtain the original object's address with appropriate type. (See the argument comm below.)

Note: f should not return floating-point NaN (Not a Number) or infinity values, since these are not handled by d01smc. If your code inadvertently does return any NaNs or infinities, d01smc is likely to produce unexpected results.

2: $boundinf$ – Nag_BoundInterval Input

On entry: indicates the kind of integration interval.

$boundinf = Nag_UpperSemiInfinite$: The interval is $[bound, + \infty)$ .
$boundinf = Nag_LowerSemiInfinite$: The interval is $(- \infty, bound]$ .
$boundinf = Nag_Infinite$: The interval is $(- \infty, + \infty)$ .

Constraint:

boundinf = Nag_UpperSemiInfinite

Nag_LowerSemiInfinite

Nag_Infinite

3: $bound$ – double Input

On entry: the finite limit of the integration interval (if present). bound is not used if

boundinf = Nag_Infinite

4: $epsabs$ – double Input

On entry: the absolute accuracy required. If epsabs is negative, the absolute value is used. See Section 7.

5: $epsrel$ – double Input

On entry: the relative accuracy required. If epsrel is negative, the absolute value is used. See Section 7.

6: $max_num_subint$ – Integer Input

On entry: the upper bound on the number of sub-intervals into which the interval of integration may be divided by the function. The more difficult the integrand, the larger max_num_subint should be.

Constraint:

max_num_subint \geq 1

7: $result$ – double * Output

On exit: the approximation to the integral

I

8: $abserr$ – double * Output

On exit: an estimate of the modulus of the absolute error, which should be an upper bound for

| I - result |

9: $qp$ – Nag_QuadProgress *

Pointer to structure of type Nag_QuadProgress with the following members:

num_subint – IntegerOutput: On exit: the actual number of sub-intervals used.

fun_count – IntegerOutput: On exit: the number of function evaluations performed by d01smc.

sub_int_beg_pts – double *Output
sub_int_end_pts – double *Output
sub_int_result – double *Output
sub_int_error – double *Output: On exit: these pointers are allocated memory internally with max_num_subint elements. If an error exit other than NE_INT_ARG_LT, NE_BAD_PARAM or NE_ALLOC_FAIL occurs, these arrays will contain information which may be useful. For details, see Section 9.
Before a subsequent call to d01smc is made, or when the information contained in these arrays is no longer useful, you should free the storage allocated by these pointers using the NAG macro NAG_FREE.

10: $comm$ – Nag_User *

Pointer to a structure of type Nag_User with the following member:

p – Pointer: On entry/exit: the pointer $comm \to p$ , of type Pointer, allows you to communicate information to and from f(). An object of the required type should be declared, e.g., a structure, and its address assigned to the pointer $comm \to p$ by means of a cast to Pointer in the calling program, e.g., comm.p = (Pointer)&s. The type Pointer is void *.

11: $fail$ – NagError * Input/Output

The NAG error argument (see Section 7 in the Introduction to the NAG Library CL Interface).

6 Error Indicators and Warnings

NE_ALLOC_FAIL: Dynamic memory allocation failed.
NE_BAD_PARAM: On entry, argument boundinf had an illegal value.
NE_INT_ARG_LT: On entry, max_num_subint must not be less than 1: $max_num_subint = ⟨ value ⟩$ .
NE_QUAD_BAD_SUBDIV: Extremely bad integrand behaviour occurs around the sub-interval $(⟨ value ⟩, ⟨ value ⟩)$ .
The same advice applies as in the case of NE_QUAD_MAX_SUBDIV.
NE_QUAD_BAD_SUBDIV_INTS: Extremely bad integrand behaviour occurs around one of the sub-intervals $(⟨ value ⟩, ⟨ value ⟩)$ or $(⟨ value ⟩, ⟨ value ⟩)$ .
The same advice applies as in the case of NE_QUAD_MAX_SUBDIV.
NE_QUAD_MAX_SUBDIV: The maximum number of subdivisions has been reached: $max_num_subint = ⟨ value ⟩$ .

The maximum number of subdivisions has been reached without the accuracy requirements being achieved. Look at the integrand in order to determine the integration difficulties. If the position of a local difficulty within the interval can be determined (e.g., a singularity of the integrand or its derivative, a peak, a discontinuity, etc.) you will probably gain from splitting up the interval at this point and calling the integrator on the sub-intervals. If necessary, another integrator, which is designed for handling the type of difficulty involved, must be used. Alternatively, consider relaxing the accuracy requirements specified by epsabs and epsrel, or increasing the value of max_num_subint.
NE_QUAD_NO_CONV: The integral is probably divergent or slowly convergent.
Please note that divergence can also occur with any error exit other than NE_INT_ARG_LT, NE_BAD_PARAM or NE_ALLOC_FAIL.
NE_QUAD_ROUNDOFF_EXTRAPL: Round-off error is detected during extrapolation.
The requested tolerance cannot be achieved, because the extrapolation does not increase the accuracy satisfactorily; the returned result is the best that can be obtained.
The same advice applies as in the case of NE_QUAD_MAX_SUBDIV.
NE_QUAD_ROUNDOFF_TOL: Round-off error prevents the requested tolerance from being achieved: $epsabs = ⟨ value ⟩$ , $epsrel = ⟨ value ⟩$ .
The error may be underestimated. Consider relaxing the accuracy requirements specified by epsabs and epsrel.

7 Accuracy

d01smc cannot guarantee, but in practice usually achieves, the following accuracy:

| I - result | \leq tol

where

tol = \max {| epsabs |, | epsrel | \times | I |}

and epsabs and epsrel are user-specified absolute and relative error tolerances. Moreover it returns the quantity abserr which, in normal circumstances, satisfies

| I - result | \leq abserr \leq tol .

8 Parallelism and Performance

d01smc is not threaded in any implementation.

9 Further Comments

The time taken by d01smc depends on the integrand and the accuracy required.

If the function fails with an error exit other than NE_INT_ARG_LT, NE_BAD_PARAM or NE_ALLOC_FAIL then you may wish to examine the contents of the structure qp. These contain the end-points of the sub-intervals used by d01smc along with the integral contributions and error estimates over the sub-intervals.

Specifically,

i = 1, 2, \dots, n

, let

r_{i}

denote the approximation to the value of the integral over the sub-interval

[a_{i}, b_{i}]

in the partition of

[a, b]

and

e_{i}

be the corresponding absolute error estimate.

Then,

\int_{a_{i}}^{b_{i}} f (x) d x ≃ r_{i}

and

result = \sum_{i = 1}^{n} r_{i}

unless the function terminates while testing for divergence of the integral (see Section 3.4.3 of Piessens et al. (1983)). In this case, result (and abserr) are taken to be the values returned from the extrapolation process. The value of