×

Dahlquist’s classical papers on stability theory. (English) Zbl 1112.65074

Dedicated to the memory of Germund Dahlquist the author describes seminal work of G. Dahlquist [Math. Scand. 4, 33–53 (1956; Zbl 0071.11803) and Proc. Symp. Appl. Math. 15, 147–158 (1963; Zbl 0123.32405)] in which he introduced criteria for convergence and stability of difference approximations to solutions of ordinary differential equations. This opened a new field of research.
The first paper considered a \(p\)-step approximation (i) \(\sum^p_{j=0}\alpha_jy_{j+n} =h\sum^p_{j=0}\beta_jf_{j+n}\), to an ordinary differential equation \(y' =f(t,y)\), \(t,y\) in \(\mathbb R\), associated polynomials (ii) \(\rho(\zeta)= \sum^p_{j=0}\alpha_j \zeta^j\), \(\sigma(\zeta)=\sum^p_{j=0}\beta_j \zeta^j\), and (iii) \(R(z)\), \(S(z)\) obtained from (ii) by the transformation \(\zeta=z+1/z-1\). The solution of (i) is called stable if the roots of \(R(z)= 0\) are in the left half plane or with multiplicity one on the imaginary axis. It was proved that for convergence of a stable system of order \(k\) that the degree \(p\) of (ii) could not exceed \(p=k+2\). If \(R(z)\) is an odd function then \(p=k+2\) with all roots on \(\text{Im}\,z=0\). If \(k\) is odd then \(p=k+1\). If \(k\) is even then \(p=k+2\) when \(\alpha_v=\alpha_{k-v}\), \(\beta_v=\beta_{k-v}\) and all roots of \(R(z)\) are on \(\text{Im}\, z=0\).
Interest in stiff systems of differential equations led to the introduction of \(A\)-stability in the second paper. The system (i) is \(A\)-stable if all solutions of (i) applied to \(y'=qy\), \(\text{Re}\,q<0\), tend to zero \(n\to\infty\). This led to the result that the order \(p\) of an \(A\)-stable method (i) could not exceed \(p=2\)., The smallest error constant occurs with the trapezoid rule. Later Dahlquist introduced \(G\)-stability which is discussed in a companion paper by J. C. Butcher [BIT 46, No. 3, 479–489 (2006; Zbl 1105.65085)]. Following the introduction of \(A\)-stability the idea of analysing stability of various discrete systems was successfully followed by many researchers, introducing appropriate polynomials. The author sketches many important conjectures and theorems in this direction. The paper includes warm reminiscences of this distinguished scientist by the author.

MSC:

65L20 Stability and convergence of numerical methods for ordinary differential equations
65-03 History of numerical analysis
65L06 Multistep, Runge-Kutta and extrapolation methods for ordinary differential equations
01A70 Biographies, obituaries, personalia, bibliographies
34A34 Nonlinear ordinary differential equations and systems
65L05 Numerical methods for initial value problems involving ordinary differential equations

Biographic References:

Dahlquist, Germund

Software:

RODAS

References:

[1] A. Abdulle, Fourth order Chebyshev methods with recurrence relation, SIAM J. Sci. Comp., 23 (2002), pp. 2042–2055. · Zbl 1009.65048 · doi:10.1137/S1064827500379549
[2] G. Birkhoff and R. S. Varga, Discretization errors for well-set Cauchy problems, 1, J. Math. Phys., 44 (1965), pp. 1–23. · Zbl 0134.13406
[3] J. Butcher, Thirty years of G-stability, this issue, DOI: 10.1007/s10543-006-0078-8. · Zbl 1105.65085
[4] G. Dahlquist, Convergence and stability in the numerical integration of ordinary differential equations, Math. Scand., 4 (1956), pp. 33–53. · Zbl 0071.11803
[5] G. Dahlquist, Stability and error bounds in the numerical integration of ordinary differential equations, Trans. Royal Inst. Technol., Stockholm, Sweden, Nr.130, p. 87, 1959. · Zbl 0085.33401
[6] J. W. Daniel and R. E. Moore, Computation and theory in ordinary differential equations, W. H. Freeman and Company, San Francisco, p. 172, 1970. · Zbl 0207.08802
[7] B. L. Ehle, High order A-stable methods for the numerical solution of systems of D.E.’s, BIT, 8 (1968), pp. 276–278. · Zbl 0176.14604 · doi:10.1007/BF01933437
[8] Y. Genin, An algebraic approach to A-stable linear multistep-multiderivative integration formulas, BIT, 14 (1974), pp. 382–406. · Zbl 0322.65037 · doi:10.1007/BF01932535
[9] E. Hairer, Symmetric linear multistep methods, this issue, DOI: 10.1007/s10543-006-0066-z. · Zbl 1149.65054
[10] E. Hairer, S. P. Nørsett, and G. Wanner, Solving Ordinary Differential Equations I, Nonstiff Problems, 2nd edn., Springer, Berlin, Heidelberg, 1993. · Zbl 0789.65048
[11] E. Hairer and G. Wanner, Solving Ordinary Differential Equations II, Stiff and Differential-Algebraic Problems, 2nd. edn., Springer, Berlin, Heidelberg, 1991, 1996. · Zbl 0729.65051
[12] P. Henrici, Discrete Variable Methods in Ordinary Differential Equations, John Wiley & Sons Inc., New York, London, Sydney, 1962. · Zbl 0112.34901
[13] W. Hundsdorfer and J. G. Verwer, Numerical Solution of Time-Dependent Advectiondiffusion-Reaction Equations, Springer, Berlin, Heidelberg, 2003. · Zbl 1030.65100
[14] R. Jeltsch, Note on A-stability of multistep multiderivative methods, BIT, 16 (1976), pp. 74–78. · Zbl 0321.65043 · doi:10.1007/BF01940779
[15] R. Jeltsch and O. Nevanlinna, Largest disk of stability of explicit Runge–Kutta methods, BIT, 18 (1978), pp. 500–502. · Zbl 0399.65051 · doi:10.1007/BF01932030
[16] R. Jeltsch and O. Nevanlinna, Stability of explicit time discretizations for solving initial value problems, Numer. Math., 37 (1981), pp. 61–91 (Corrigendum: Numer. Math., 39 (1982), p. 155). · Zbl 0457.65054
[17] S. P. Nørsett, C-polynomials for rational approximations to the exponential function, Numer. Math., 25 (1975), pp. 39–56. · Zbl 0299.65010 · doi:10.1007/BF01419527
[18] M. Reimer, Zur Theorie der linearen Differenzenformeln, Math. Z., 95 (1967), pp. 373–402. · Zbl 0143.37902 · doi:10.1007/BF01111689
[19] G. Wanner, E. Hairer, and S. P. Nørsett, Order stars and stability theorems, BIT, 18 (1978), pp. 475–489. · Zbl 0444.65039 · doi:10.1007/BF01932026
[20] G. Wanner, E. Hairer, and S. P. Nørsett, When I-stability implies A-stability, BIT, 18 (1978), p. 503. · Zbl 0444.65040 · doi:10.1007/BF01932031
This reference list is based on information provided by the publisher or from digital mathematics libraries. Its items are heuristically matched to zbMATH identifiers and may contain data conversion errors. In some cases that data have been complemented/enhanced by data from zbMATH Open. This attempts to reflect the references listed in the original paper as accurately as possible without claiming completeness or a perfect matching.