Advertisement

Numerische Mathematik

, Volume 55, Issue 6, pp 735–745 | Cite as

On the construction of multi-dimensional embedded cubature formulae

  • Ronald Cools
  • Ann Haegemans
Article

Summary

In order to compute an integralI[f], one needs at least two cubature formulaeQ j ,j∈{1, 2}. |Q1[f]−Q2[f]| can be used as an error estimate for the less precise cubature formula. In order to reduce the amount of work, one can try to reuse some of the function evaluations needed forQ1, inQ2. The easiest way to construct embedded cubature formulae is: start with a high degree formulaQ1, drop (at least) one knot and calculate the weights such that the new formulaQ2 is exact for as much monomials as possible. We describe how such embedded formulae with positive weights can be found. The disadvantage of such embedded cubature formulae is that there is in general a large difference in the degree of exactness of the two formulae. In this paper we will explain how the high degree formula can be chosen to obtain an embedded pair of cubature formulae of degrees 2m+1/2m−1. The method works for all regionsΩ⊂ℝ n ,n≧2. We will also show the influence of structure on the cubature formulae.

Subject Classifications

AMS(MOS): 65D32 41A55 41A63 CR: G.1.4 

Preview

Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.

References

  1. 1.
    Berntsen, J., Espelid, T.: On the use of Gauss quadrature in adaptive integration schemes. BIT24, 239–242 (1984)Google Scholar
  2. 2.
    Berntsen, I., Espelid, T.: On the construction of higher degree three-dimensional embedded integration rules. SIAM J. Numer. Anal.25, 222–234 (1988)CrossRefGoogle Scholar
  3. 3.
    Cools, R., Haegemans, A.: Optimal addition of knots to cubature formulae for planar regions. Numer. Math.49, 269–274 (1986)Google Scholar
  4. 4.
    Cools, R., Haegemans, A.: Construction of sequences of embedded cubature formulae for circular symmetric planar regions. In: Keast, P., Fairweather, G. (eds.) Numerical Integration, pp. 165–172. Dordrecht: Reidel 1987Google Scholar
  5. 5.
    Cools, R., Haegemans, A.: A lower bound for the number of function evaluations in an error estimate for numerical integration. Report TW111 K.U. Leuven, Department of Computer Science 1988Google Scholar
  6. 6.
    Franke, R.: Orthogonal polynomials and approximate multiple integration. SIAM J. Numer. Anal.8, 757–766 (1971)CrossRefGoogle Scholar
  7. 7.
    Genz, A.C., Malik, A.A.: An imbedded family of fully symmetric numerical integration rules. SIAM J. Numer. Anal.20, 580–588 (1983)CrossRefGoogle Scholar
  8. 8.
    Gröbner, W.: Moderne Algebraische Geometrie. Wien: Springer 1949Google Scholar
  9. 9.
    Laurie, D.: Practical error estimates in numerical integration. J. Comput. Appl. Math.12 & 13, 425–431 (1985)Google Scholar
  10. 10.
    Möller, H.M.: Polynomideale und Kubaturformeln. Dissertation, Universität Dortmund 1973Google Scholar
  11. 11.
    Möller, H.M.: Kubaturformeln mit minimaler, Knotenzahl. Numer Math.25, 185–200 (1976)Google Scholar
  12. 12.
    Möller, H.M.: On the construction of cubature formulae with few nodes using Gröbner bases. In: Keast, P., Fairweather, G. (eds). Numerical integraton, pp. 177–192. Dordrecht: Reidel 1987Google Scholar
  13. 13.
    Piessens, R., Haegemans, A.: Cubature formulas of degree nine for symmetric planar regions. Math. Comput.29, pp. 810–815 (1975)Google Scholar
  14. 14.
    Rabinowitz, P., Kautsky, J., Elhay, S., Butcher, J.: On sequences of imbedded integration rules. In: Keast, P., Fairweather, G. (eds.) Numerical integration, pp. 113–139. Dordrecht: Reidel 1987Google Scholar
  15. 15.
    Stroud, A.H.: Approximate calculation of multiple integrals. Englewood Cliffs, N.J.: Prentice Hall 1971Google Scholar

Copyright information

© Springer-Verlag 1989

Authors and Affiliations

  • Ronald Cools
    • 1
  • Ann Haegemans
    • 1
  1. 1.Department of Computer ScienceKatholieke Universiteit LeuvenHeverleeBelgium

Personalised recommendations