Abstract
Efficient general quadrature formulas with nodes of arbitrary multiplicity are constructed for numerical integration of rapidly oscillating functions. The quadrature weights of these formulas are derived in explicit form in terms of easily evaluated integrals of products of a rapidly oscillating function and elementary basis functions and also in terms of elements of the inverse of the Vandermonde matrix. Error bounds are established for the quadrature formulas, which depend both on the integration increment and on the oscillation parameters. Necessary conditions are proved when the relative error is bounded and tends to zero with the increase of the oscillation parameters. Two-sided interpolation polynomials are applied to obtain easily computed posterior error bounds for quadrature formulas.
Similar content being viewed by others
References
N. S. Bakhvalov, N. P. Zhidkov, and G. M. Kobel'kov, Numerical Methods [in Russian], Moscow (1987).
A. F. Kalaida, Matrix Algorithms for Approximate and Numerical Integration [in Russian], Kiev (1982). Unpublished manuscript, VINITI 07.9.82, No. 3672-82.
A. F. Kalaida, High-Rank Collocational Approximation and Matrix Algorithms for Numerical Differentiation and Integration [in Russian], Kiev (1982). Unpublished manuscript, VINITI 12.31.82, No. 5662-82.
A. F. Kalaida, "On the error of numerical differentiation formulas," Vychisl. Prikl. Mat., No. 54, 16–26 (1984).
A. F. Kalaida, "On two-sided approximations of functions by polynomials," Vychisl. Prikl. Mat., No. 67, 31–37 (1989).
Additional information
Kiev University. Translated from Vychislitel'naya i Prikladnaya Matematika, No. 75, pp. 12–20, 1991.
Rights and permissions
About this article
Cite this article
Kalaida, A.F. Matrix numerical integration algorithms for rapidly oscillating functions. J Math Sci 72, 3053–3060 (1994). https://doi.org/10.1007/BF01259470
Received:
Issue Date:
DOI: https://doi.org/10.1007/BF01259470