Abstract
When solving systems of nonlinear equations, generally the first step is to linearize them. When using interval methods, one usually preconditions the linear equations by multiplying by a real (i.e. non-interval) matrix which approximates the inverse of the center of the interval Jacobian. In this paper, we show that it is better to precondition the original nonlinear equations rather than the linearized ones.
Zusammenfassung
Bei der Lösung nichtlinearer Gleichungssysteme wird im allgemeinen zuerst linearisiert. Wenn Intervall-Methoden benützt werden, werden die linearen Gleichungen üblicherweise durch Multiplikation mit geeigneten reellen Zahlen konditioniert. Wir zeigen, daß eine Präkonditionierung der ursprünglichen, nicht linearisierten Gleichungen besser ist.
Similar content being viewed by others
References
Hansen, E. R.: Global optimization using interval analysis. New York: Marcel Dekker, 1992.
Hansen, E. R., Smith, R. R.: Interval arithmetic in matrix computations, part 2. SIAM J. Numer. Anal.4, 1–9 (1967).
Hu, C.: Optimal preconditioners for the interval Newton method. Ph.D. Dissertation, University of Southwestern Louisiana, 1990.
Kearfott, R. B.: Preconditioners for the interval Gauss-Seidel method, SIAM J. Numer. Anal.27, 804–822 (1990).
Moore, R. E.: Methods and applications of interval analysis. Philadelphia: SIAM, 1979.
Neumaier, A.: Interval methods for systems of equations. Cambridge: Cambridge University Press, 1990.
Novoa, M.: Linear programming preconditioners for the interval Gauss-Seidel method. Ph.D. Dissertation, University of Southwestern Louisiana, 1992.
Ratschek, H., Rokne, J.: Computer methods for the range of functions. Chichester: Ellis Horwood, 1984.
Author information
Authors and Affiliations
Rights and permissions
About this article
Cite this article
Hansen, E.R. Preconditioning linearized equations. Computing 58, 187–196 (1997). https://doi.org/10.1007/BF02684439
Received:
Revised:
Issue Date:
DOI: https://doi.org/10.1007/BF02684439