Advertisement

Computing

, Volume 51, Issue 2, pp 135–150 | Cite as

On the use of parallel processors for implicit Runge-Kutta methods

  • G. J. Cooper
  • R. Vignesvaran
Article

Abstract

An iteration scheme, for solving the non-linear equations arising in the implementation of implicit Runge-Kutta methods, is proposed. This scheme is particularly suitable for parallel computation and can be applied to any method which has a coefficient matrixA with all eigenvalues real (and positive). For such methods, the efficiency of a modified Newton scheme may often be improved by the use of a similarity transformation ofA but, even when this is the case, the proposed scheme can have advantages for parallel computation. Numerical results illustrate this. The new scheme converges in a finite number of iterations when applied to linear systems of differential equations, achieving this by using the nilpotency of a strictly lower triangular matrixS−1AS — Λ, with Λ a diagonal matrix. The scheme reduces to the modified Newton scheme whenS−1AS is diagonal.A convergence result is obtained which is applicable to nonlinear stiff systems.

AMS Subject Classification

AMS (MOS) 65 L 05 

Key words

Implementation implicit methods Runge-Kutta parallel processing 

Parallele Verarbeitung für implizite Runge-Kutta-Verfahren

Zusammenfassung

Wir schlagen ein Iterationsverfahren für die Lösung der nicht-linearen Gleichungen bei der Implementierung von impliziten Runge-Kutta-Verfahren vor, das besonders geeignet für die Parallelverarbeitung ist und immer angewandt werden kann, wenn die KoeffizientenmatrixA nur reelle (und positive) Eigenwerte hat. Für solche Verfahren kann man zwar die Effizienz eines modifizierten Newton-Verfahrens oft auch mit Hilfe einer Ähnlichkeitstransformation vonA verbessern, aber sogar dann hat unser vorgeschlagenes Verfahren Vorteile für die Parallelverarbeitung, wie numerische Beispiele zeigen. Das neue Vorgehen konvergiert für lineare Differentialgleichungssysteme in endlich vielen Schritten wegen der Nilpotenz einer strengen unteren DreiecksmatrixS−1AS — Λ und reduziert sich auf das modifizierte Newton-Verfahren wennS−1 eine Diagonalmatrix ist. Unser Konvergenzresultat ist auf nichtlineare steife Systeme anwendbar.

Preview

Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.

References

  1. [1]
    Alexander, R.: The modified Newton method in the solution of stiff ordinary differential equations. Math. Comp.57, 673–701 (1991).Google Scholar
  2. [2]
    Bickart, T. A.: An efficient solution process for implicit Runge-Kutta methods. SIAM J. Numer. Anal.14, 1022–1027 (1977).Google Scholar
  3. [3]
    Burrage, K.: A special family of Runge-Kutta methods for solving stiff differential equatons. BIT18, 22–41 (1978).Google Scholar
  4. [4]
    Burrage, K.: Efficiently implementable algebraically stable Runge-Kutta methods. SIAM J. Numer. Anal.19, 245–258 (1982).Google Scholar
  5. [5]
    Burrage, K., Butcher, J. C., Chipman, F. H.: An implementation of singly-implicit Runge-Kutta methods. BIT20, 326–340 (1980).Google Scholar
  6. [6]
    Butcher, J. C.: On the implementation of implicit Runge-Kutta methods. BIT16, 237–240 (1976).Google Scholar
  7. [7]
    Butcher, J. C.: Some implementation schemes for implicit Runge-Kutta methods. Proc. Dundee Conf. on Numerical Analysis, Lecture notes in Mathematics pp. 12–24. Berlin Heidelberg New York: Springer 1979.Google Scholar
  8. [8]
    Cash, J. R.: On a class of implicit Runge-Kutta procedures. J. Inst. Math. Appl.19, 455–470 (1977).Google Scholar
  9. [9]
    Chipman, F. H.: The implementation of Runge-Kutta implicit processes. BIT13, 391–393 (1973).Google Scholar
  10. [10]
    Collings, A. G., Tee, G. J.: An analysis of Euler and implicit Runge-Kutta numerical integration schemes for structural dynamic problems. Proc. Sixth Australasian Conf. on the Mechanics of Structures and Materials 1977,1, 147–154.Google Scholar
  11. [11]
    Cooper, G. J.: On the implementation of singly implicit Runge-Kutta methods. Math. Comp.57, 663–672 (1991).Google Scholar
  12. [12]
    Cooper, G. J.: On the existence of solutions for algebraically stable Runge-Kutta methods. IMA J. Numer. Anal.6, 325–330 (1986).Google Scholar
  13. [13]
    Cooper, G. J., Butcher, J. C.: An iteration scheme for implicit Runge-Kutta methods. IMA J. Numer. Anal.3, 127–140 (1983).Google Scholar
  14. [14]
    Cooper, G. J., Sayfy, A.: Semi-explicitA-stable Runge-Kutta methods. Math. Comp.33, 541–556 (1979).Google Scholar
  15. [15]
    Cooper, G. J., Vignesvaran, R.: A scheme for the implementation of implicit Runge-Kutta methods. Computing45, 321–332 (1990).Google Scholar
  16. [16]
    Cooper, G. J., Vignesvaran, R.: Some schemes for the implementation of implicit Runge-Kutta methods. J. Comp. Appl. Math.45, 213–225 (1993).Google Scholar
  17. [17]
    Crouzeix, M., Hundsdorfer, W. H., Spijker, M. N.: On the existence of solutions to the algebraic equations in implicit Runge-Kutta methods. BIT23, 84–91 (1983).Google Scholar
  18. [18]
    Davis, H. T.: Introduction to non-linear differential and integral equations. New York: Dover 1962.Google Scholar
  19. [19]
    Dekker, K.: Error bounds for the solution to the algebraic equations in Runge-Kutta methods. BIT24, 347–356 (1984).Google Scholar
  20. [20]
    Dekker, K., Verwer, J. G.: Stability of Runge-Kutta methods for stiff nonlinear differential equations. Amsterdam. North-Holland, 1984.Google Scholar
  21. [21]
    Enright, W. H.: Improving the efficiency of matrix operations in the numerical solution of stiff ordinary differential equations. ACM Trans. Math. Software4, 127–136 (1978).Google Scholar
  22. [22]
    Frank, R., Ueberhuber, C. W.: Iterated defect correction for the efficient solution of stiff systems of ordinary differential equations. BIT17, 146–159 (1977).Google Scholar
  23. [23]
    Gear, C. W.: The potential for parallelsim in ordinary differential equations. Tech. Rep. UIUC-DCS-R-86-1246, Computer Science Dept., Univ. of Illinois 1986.Google Scholar
  24. [24]
    Gear, C. W.: Parallel methods in ordinary differential equations, Tech. Rep. UIUC-DCS-R-87-1396. Computer Science Dept., Univ. of Illinois, 1987.Google Scholar
  25. [25]
    Gear, C. W.: The automatic integration of stiff ordinary differential equations. Proc. IFIP Congress 1968, 187–193.Google Scholar
  26. [26]
    Hairer, E., Norsett, S. P. Wanner, G.: Solving ordinary differential equations I. Nonstiff problems. Berlin Heidelberg New York Tokyo: Springer 1987.Google Scholar
  27. [27]
    Hairer, E., Wanner, G.: Algebraically stable and implementalbe Runge-Kutta methods of high order. SIAM J. Numer. Anal.18, 1098–1108 (1981).Google Scholar
  28. [28]
    van der Houwen, P. J., Sommeijer, B. P.: Iterated Runge-Kutta methods on parallel computers. SIAM J. Sci. Comput.12, 1000–1028 (1991).Google Scholar
  29. [29]
    Iserles, A., Norsett, S. P.: On the theory of parallel Runge-Kutta methods. IMA J. Numer. Anal.10, 463–488 (1990).Google Scholar
  30. [30]
    Jackson, K. R., Norsett, S. P.: The potential for parallelism in Runge-Kutta methods. Part 1: RK formulas in standard form, Tech. Rep. 239/90, Computer Science Dept., Univ. of Toronto 1990.Google Scholar
  31. [31]
    Karakashian, O. A., Rust, W.: On the parallel implementation of implicit Runge-Kutta methods. SIAM J. Sci. Stat. Comput.9, 1085–1090 (1988).Google Scholar
  32. [32]
    Keeling, S. L.: On implicit Runge-Kutta methods with a stability function having real poles. BIT29, 91–109 (1989).Google Scholar
  33. [33]
    Kraaijevanger, J. F. B. M., Schneid, J.: On the unique solvability of the Runge-Kutta equations. Numer. Anal. Rep. 90-6, Mathematics Dept., Massachusetts Institute of Technology 1990.Google Scholar
  34. [34]
    Kreiss, H. O.: Difference methods for stiff ordinary differential equations. SIAM J. Numer. Anal.15, 21–58 (1978).Google Scholar
  35. [35]
    Norsett, S. P.: Runge-Kutta methods with a multiple real eigenvalue only. BIT16, 388–393 (1976).Google Scholar
  36. [36]
    Norsett, S. P., Simonsen, H. H.: Aspects of parallel Runge-Kutta methods. In: Numerical methods for ordinary differential equations. Proc. L'Aquila 1987, pp. 103–107. Berlin Heidelberg New York Tokyo: Springer 1989.Google Scholar
  37. [37]
    Wanner, G., Hairer, E., Norsett, S. P.: Order stars and stability theorems. BIT18, 475–489 (1978).Google Scholar

Copyright information

© Springer-Verlag 1993

Authors and Affiliations

  • G. J. Cooper
    • 1
  • R. Vignesvaran
    • 2
  1. 1.School of Mathematical and Physical SciencesUniversity of SussexBrightonEngland
  2. 2.Department of MathematicsEastern UniversityChenkaladySri Lanka

Personalised recommendations