pith. sign in

arxiv: 2505.15099 · v4 · submitted 2025-05-21 · 🧮 math.NA · cs.NA

A Stiff Order Condition Theory for Runge-Kutta Methods Applied to Semilinear ODEs

Steven B. Roberts , David Shirokoff , Abhijit Biswas , Benjamin Seibold This is my paper

Pith reviewed 2026-05-22 14:32 UTC · model grok-4.3

classification 🧮 math.NA cs.NA
keywords Runge-Kutta methodssemilinear ODEsorder conditionsstiff problemsorder reductionrooted treesglobal error bounds
0
0 comments X

The pith

Weaker orthogonality conditions suffice for Runge-Kutta methods to achieve full order on semilinear stiff ODEs

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

This paper shows that classical order conditions for Runge-Kutta methods can be relaxed for semilinear ODEs that combine a stiff linear term with a non-stiff nonlinear term. The relaxed conditions take the form of orthogonality relations that are counted using rooted trees. A sympathetic reader would care because these weaker conditions prevent the loss of accuracy known as order reduction even when the time step is not small relative to the stiffness. The work also establishes global error bounds that stay valid no matter how large the stiffness parameter becomes.

Core claim

For the broad class of semilinear ODEs consisting of a stiff linear term and non-stiff nonlinear term, weaker conditions than the classical ones suffice to prevent order reduction. These new semilinear order conditions are formulated in terms of orthogonality relations and can be enumerated by rooted trees. Global error bounds are proved that hold uniformly with respect to the stiffness of the linear term.

What carries the argument

Semilinear order conditions formulated as orthogonality relations that are enumerated by rooted trees

Load-bearing premise

The ODE must be semilinear with a stiff linear term whose stiffness can grow arbitrarily large while the nonlinear term remains non-stiff.

What would settle it

A numerical test on a semilinear problem with increasing stiffness where a Runge-Kutta method satisfying the orthogonality relations still shows order reduction in the global error.

read the original abstract

Classical convergence theory of Runge-Kutta methods assumes that the time step is small relative to the Lipschitz constant of the ordinary differential equation (ODE). For stiff problems, that assumption is often violated, and a problematic degradation in accuracy, known as order reduction, can arise. Methods with high stage order, e.g., Gauss-Legendre and Radau, are known to avoid order reduction, but they must be fully implicit. For the broad class of semilinear ODEs, which consist of a stiff linear term and non-stiff nonlinear term, we show that weaker conditions suffice. Our new semilinear order conditions are formulated in terms of orthogonality relations and can be enumerated by rooted trees. Finally, we prove global error bounds that hold uniformly with respect to stiffness of the linear term.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

2 major / 2 minor

Summary. The paper develops a stiff order condition theory for Runge-Kutta methods on semilinear ODEs of the form y' = Ly + f(y), where L is a stiff linear operator and f is non-stiff and nonlinear. It derives new order conditions expressed as orthogonality relations that can be enumerated using rooted trees, shows these are weaker than classical conditions, and proves global error bounds that remain uniform with respect to the stiffness of L.

Significance. If the derivations and proofs hold, the result is significant for numerical ODEs: it identifies a broad class of methods that avoid order reduction on semilinear stiff problems without requiring the high stage order of fully implicit methods such as Gauss or Radau. The rooted-tree enumeration and uniform-in-stiffness bounds constitute a clean extension of Butcher-series techniques to this setting and could guide the design of more efficient integrators for applications like reaction-diffusion equations.

major comments (2)
  1. [§4, Theorem 4.3] §4, Theorem 4.3: the uniform global error bound is stated to hold for arbitrary stiffness, yet the proof sketch invokes a bound on the nonlinear Lipschitz constant that is independent of the linear operator norm; this assumption is load-bearing for uniformity and should be stated explicitly as a hypothesis rather than left implicit in the semilinear splitting.
  2. [§3.1, Eq. (3.4)] §3.1, Eq. (3.4): the orthogonality relations are derived by enumerating trees up to a given order; it is not immediately clear whether the same tree set yields the claimed order for all Runge-Kutta methods or only for those already satisfying the classical order conditions, which affects the practical utility of the new conditions.
minor comments (2)
  1. [Introduction] The notation for the linear operator L and its resolvent could be introduced earlier and used consistently to improve readability of the tree-based derivations.
  2. [§3] Figure 2 (tree enumeration) would benefit from an explicit legend indicating which trees correspond to the new semilinear conditions versus the classical ones.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the careful reading of our manuscript and the constructive comments. We address each major comment below and have incorporated revisions to improve clarity.

read point-by-point responses

  1. Referee: §4, Theorem 4.3: the uniform global error bound is stated to hold for arbitrary stiffness, yet the proof sketch invokes a bound on the nonlinear Lipschitz constant that is independent of the linear operator norm; this assumption is load-bearing for uniformity and should be stated explicitly as a hypothesis rather than left implicit in the semilinear splitting.

    Authors: We agree that the independence of the nonlinear Lipschitz constant from the norm of the linear operator is essential to the uniformity claim and should be stated explicitly. In the revised manuscript we will add this as a formal hypothesis in the statement of Theorem 4.3 and will expand the proof sketch in Section 4 to reference the hypothesis directly. revision: yes

  2. Referee: §3.1, Eq. (3.4): the orthogonality relations are derived by enumerating trees up to a given order; it is not immediately clear whether the same tree set yields the claimed order for all Runge-Kutta methods or only for those already satisfying the classical order conditions, which affects the practical utility of the new conditions.

    Authors: The rooted-tree enumeration in Section 3.1 yields the complete set of semilinear order conditions that are sufficient by themselves for the uniform error bounds. These conditions are strictly weaker than the classical Butcher conditions and apply to any Runge-Kutta method that satisfies the listed orthogonality relations, without requiring prior satisfaction of the full classical set. We will revise the paragraph following Equation (3.4) to state this explicitly and to emphasize the relaxation relative to classical theory. revision: yes

Circularity Check

0 steps flagged

No significant circularity identified

full rationale

The paper adapts standard Butcher-series and rooted-tree techniques to semilinear ODEs by splitting into a stiff linear term and non-stiff nonlinear term, then derives orthogonality-based order conditions directly from that splitting. Global error bounds uniform in stiffness follow from these conditions via standard stability and consistency arguments. No step reduces by construction to a fitted parameter, self-definition, or load-bearing self-citation; the derivation remains independent of the target result and uses externally verifiable mathematical steps.

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 0 invented entities

The central claim rests on the semilinear splitting of the ODE and the validity of the new orthogonality conditions; no free parameters or invented entities are introduced in the abstract.

axioms (1)
  • domain assumption The ODE is semilinear, consisting of a stiff linear term and a non-stiff nonlinear term.
    This splitting is the defining assumption that allows weaker conditions than the fully stiff case.

pith-pipeline@v0.9.0 · 5675 in / 1250 out tokens · 38404 ms · 2026-05-22T14:32:16.586482+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

39 extracted references · 39 canonical work pages

  1. [1]

    Mathematics of Computation28(125), 145–162 (1974)

    Prothero, A., Robinson, A.: On the stability and accuracy of one-step meth- ods for solving stiff systems of ordinary differential equations. Mathematics of Computation28(125), 145–162 (1974)

    work page 1974

  2. [2]

    SIAM Journal on Numerical Analysis18(5), 753–780 (1981) https://doi.org/10.1137/ 0718051 22

    Frank, R., Schneid, J., Ueberhuber, C.W.: The concept of B-convergence. SIAM Journal on Numerical Analysis18(5), 753–780 (1981) https://doi.org/10.1137/ 0718051 22

    work page 1981

  3. [3]

    CWI monographs2(1984)

    Dekker, K., Verwer, J.G.: Stability of Runge–Kutta methods for stiff nonlinear differential equations. CWI monographs2(1984)

    work page 1984

  4. [4]

    Computing 41(3), 219–235 (1989) https://doi.org/10.1007/BF02259094

    Scholz, S.: Order barriers for the B-convergence of ROW methods. Computing 41(3), 219–235 (1989) https://doi.org/10.1007/BF02259094

    work page doi:10.1007/bf02259094 1989

  5. [5]

    SIAM Journal on Numerical Analysis22(3), 515–534 (1985) https://doi.org/10.1137/0722031

    Frank, R., Schneid, J., Ueberhuber, C.W.: Order results for implicit Runge–Kutta methods applied to stiff systems. SIAM Journal on Numerical Analysis22(3), 515–534 (1985) https://doi.org/10.1137/0722031

    work page doi:10.1137/0722031 1985

  6. [6]

    BIT Numerical Mathematics27(1), 62–71 (1987) https://doi.org/10.1007/BF01937355

    Burrage, K., Hundsdorfer, W.H.: The order of B-convergence of algebraically stable Runge–Kutta methods. BIT Numerical Mathematics27(1), 62–71 (1987) https://doi.org/10.1007/BF01937355

    work page doi:10.1007/bf01937355 1987

  7. [7]

    Springer Series in Computational Mathematics, vol

    Hairer, E., Wanner, G.: Solving Ordinary Differential Equations II: Stiff and Differential-Algebraic Problems, 2nd edn. Springer Series in Computational Mathematics, vol. 14. Springer, Berlin, Heidelberg (1996)

    work page 1996

  8. [8]

    Computing36(1), 17–34 (1986) https://doi.org/10.1007/ BF02238189

    Burrage, K., Hundsdorfer, W.H., Verwer, J.G.: A study of B-convergence of Runge–Kutta methods. Computing36(1), 17–34 (1986) https://doi.org/10.1007/ BF02238189

    work page 1986

  9. [9]

    Applied Numerical Mathematics9(2), 91–109 (1992) https:// doi.org/10.1016/0168-9274(92)90008-2

    Auzinger, W., Frank, R., Kirlinger, G.: An extension of B-convergence for Runge– Kutta methods. Applied Numerical Mathematics9(2), 91–109 (1992) https:// doi.org/10.1016/0168-9274(92)90008-2

    work page doi:10.1016/0168-9274(92)90008-2 1992

  10. [10]

    BIT Numerical Mathematics40(4), 611–639 (2000) https://doi.org/10.1023/A:1022332200092

    Calvo, M., Gonz´ alez-Pinto, S., Montijano, J.I.: Runge–Kutta methods for the numerical solution of stiff semilinear systems. BIT Numerical Mathematics40(4), 611–639 (2000) https://doi.org/10.1023/A:1022332200092

    work page doi:10.1023/a:1022332200092 2000

  11. [11]

    BIT Numerical Mathematics27(2), 264–281 (1987) https://doi.org/10

    Strehmel, K., Weiner, R.: B-convergence results for linearly implicit one step methods. BIT Numerical Mathematics27(2), 264–281 (1987) https://doi.org/10. 1007/BF01934189

    work page 1987

  12. [12]

    Computational Mathematics and Mathematical Physics43(9), 1320–1330 (2003)

    Skvortsov, L.M.: Accuracy of Runge–Kutta methods applied to stiff problems. Computational Mathematics and Mathematical Physics43(9), 1320–1330 (2003)

    work page 2003

  13. [13]

    Mathematical Models and Computer Simulations2(6), 800–811 (2010) https://doi.org/10.1134/S2070048210060165

    Skvortsov, L.M.: Model equations for accuracy investigation of Runge–Kutta methods. Mathematical Models and Computer Simulations2(6), 800–811 (2010) https://doi.org/10.1134/S2070048210060165

    work page doi:10.1134/s2070048210060165 2010

  14. [14]

    Journal of Computational Physics523, 113627 (2025) https://doi.org/10.1016/j.jcp.2024

    Cai, Y., Wan, J., Kareem, A.: On convergence of implicit Runge-Kutta methods for the incompressible Navier-Stokes equations with unsteady inflow. Journal of Computational Physics523, 113627 (2025) https://doi.org/10.1016/j.jcp.2024. 113627

    work page doi:10.1016/j.jcp.2024 2025

  15. [15]

    SIAM Journal on Numerical Analysis43(3), 1069– 1090 (2005) https://doi.org/10.1137/040611434

    Hochbruck, M., Ostermann, A.: Explicit exponential Runge–Kutta methods for 23 semilinear parabolic problems. SIAM Journal on Numerical Analysis43(3), 1069– 1090 (2005) https://doi.org/10.1137/040611434

    work page doi:10.1137/040611434 2005

  16. [16]

    SIAM Journal on Numerical Analysis51(6), 3431–3445 (2013) https://doi.org/10.1137/130920204

    Luan, V.T., Ostermann, A.: Exponential B-series: The stiff case. SIAM Journal on Numerical Analysis51(6), 3431–3445 (2013) https://doi.org/10.1137/130920204

    work page doi:10.1137/130920204 2013

  17. [17]

    Numerische Mathematik145(3), 553–580 (2020) https://doi.org/10.1007/s00211-020-01120-4

    Hochbruck, M., Leibold, J., Ostermann, A.: On the convergence of Lawson meth- ods for semilinear stiff problems. Numerische Mathematik145(3), 553–580 (2020) https://doi.org/10.1007/s00211-020-01120-4

    work page doi:10.1007/s00211-020-01120-4 2020

  18. [18]

    BIT Numerical Mathematics56(4), 1303–1316 (2016) https://doi.org/ 10.1007/s10543-016-0604-2

    Hansen, E., Ostermann, A.: High-order splitting schemes for semilinear evolution equations. BIT Numerical Mathematics56(4), 1303–1316 (2016) https://doi.org/ 10.1007/s10543-016-0604-2

    work page doi:10.1007/s10543-016-0604-2 2016

  19. [19]

    part 1: Dirichlet boundary conditions

    Einkemmer, L., Ostermann, A.: Overcoming order reduction in diffusion-reaction splitting. part 1: Dirichlet boundary conditions. SIAM Journal on Scientific Computing37(3), 1577–1592 (2015) https://doi.org/10.1137/140994204

    work page doi:10.1137/140994204 2015

  20. [20]

    part 2: Oblique boundary conditions

    Einkemmer, L., Ostermann, A.: Overcoming order reduction in diffusion-reaction splitting. part 2: Oblique boundary conditions. SIAM Journal on Scientific Computing38(6), 3741–3757 (2016) https://doi.org/10.1137/16M1056250

    work page doi:10.1137/16m1056250 2016

  21. [21]

    IMA Journal of Numerical Analysis15(4), 555–583 (1995) https://doi.org/10.1093/imanum/15.4.555

    Lubich, C., Ostermann, A.: Linearly implicit time discretization of non-linear parabolic equations. IMA Journal of Numerical Analysis15(4), 555–583 (1995) https://doi.org/10.1093/imanum/15.4.555

    work page doi:10.1093/imanum/15.4.555 1995

  22. [22]

    In: Sherwin, S.J., Moxey, D., Peir´ o, J., Vincent, P.E., Schwab, C

    Ketcheson, D.I., Seibold, B., Shirokoff, D., Zhou, D.: DIRK schemes with high weak stage order. In: Sherwin, S.J., Moxey, D., Peir´ o, J., Vincent, P.E., Schwab, C. (eds.) Spectral and High Order Methods for Partial Differential Equations ICOSAHOM 2018, pp. 453–463. Springer, Cham (2020)

    work page 2018

  23. [23]

    SIAM Journal on Numerical Analysis24(2), 391–406 (1987) https://doi.org/10.1137/0724030

    Albrecht, P.: A new theoretical approach to Runge–Kutta methods. SIAM Journal on Numerical Analysis24(2), 391–406 (1987) https://doi.org/10.1137/0724030

    work page doi:10.1137/0724030 1987

  24. [24]

    SIAM Journal on Numerical Analysis33(5), 1712–1735 (1996) https://doi.org/10.1137/S0036142994260872

    Albrecht, P.: The Runge–Kutta theory in a nutshell. SIAM Journal on Numerical Analysis33(5), 1712–1735 (1996) https://doi.org/10.1137/S0036142994260872

    work page doi:10.1137/s0036142994260872 1996

  25. [25]

    SIAM Journal on Scientific Computing31(3), 1926–1945 (2009)

    Boscarino, S., Russo, G.: On a class of uniformly accurate IMEX Runge-Kutta schemes and applications to hyperbolic systems with relaxation. SIAM Journal on Scientific Computing31(3), 1926–1945 (2009)

    work page 1926

  26. [26]

    Hu, J., Shu, R.: Uniform accuracy of implicit-explicit Runge-Kutta (IMEX-RK) schemes for hyperbolic systems with relaxation. Math. Comp.94, 209–240 (2025)

    work page 2025

  27. [27]

    Journal of Computational and Applied Mathematics255, 417–431 (2014) 24

    Luan, V.T., Ostermann, A.: Exponential Rosenbrock methods of order five — construction, analysis and numerical comparisons. Journal of Computational and Applied Mathematics255, 417–431 (2014) 24

    work page 2014

  28. [28]

    Journal of Computational and Applied Mathematics 12-13, 475–489 (1985) https://doi.org/10.1016/0377-0427(85)90041-X

    Nevanlinna, O.: Matrix valued versions of a result of von Neumann with an appli- cation to time discretization. Journal of Computational and Applied Mathematics 12-13, 475–489 (1985) https://doi.org/10.1016/0377-0427(85)90041-X

    work page doi:10.1016/0377-0427(85)90041-x 1985

  29. [29]

    Springer, Cham (2011)

    Taylor, M.E.: Partial Differential Equations I, 2nd edn. Springer, Cham (2011)

    work page 2011

  30. [30]

    Journal of Computational and Applied Mathematics 262, 105–114 (2014) https://doi.org/10.1016/j.cam.2013.09.062

    Rang, J.: An analysis of the Prothero–Robinson example for constructing new DIRK and ROW methods. Journal of Computational and Applied Mathematics 262, 105–114 (2014) https://doi.org/10.1016/j.cam.2013.09.062

    work page doi:10.1016/j.cam.2013.09.062 2014

  31. [31]

    Mathematics of Computation 59(200), 403–420 (1992) https://doi.org/10.1090/s0025-5718-1992-1142285-6

    Ostermann, A., Roche, M.: Runge–Kutta methods for partial differential equations and fractional orders of convergence. Mathematics of Computation 59(200), 403–420 (1992) https://doi.org/10.1090/s0025-5718-1992-1142285-6

    work page doi:10.1090/s0025-5718-1992-1142285-6 1992

  32. [32]

    Rosales, R.R., Seibold, B., Shirokoff, D., Zhou, D.: Spatial manifestations of order reduction in Runge-Kutta methods for initial boundary value problems. Commun. Math. Sci.22(3), 613–653 (2024) https://doi.org/10.4310/CMS.2024.v22.n3.a2

    work page doi:10.4310/cms.2024.v22.n3.a2 2024

  33. [33]

    SIAM Journal on Numerical Analysis62(1), 48–72 (2024) https://doi.org/10.1137/22M1483943

    Biswas, A., Ketcheson, D.I., Seibold, B., Shirokoff, D.: Algebraic structure of the weak stage order conditions for Runge–Kutta methods. SIAM Journal on Numerical Analysis62(1), 48–72 (2024) https://doi.org/10.1137/22M1483943

    work page doi:10.1137/22m1483943 2024

  34. [34]

    Communications in Applied Mathematics and Computational Science18(1), 1–28 (2023)

    Biswas, A., Ketcheson, D.I., Seibold, B., Shirokoff, D.: Design of DIRK schemes with high weak stage order. Communications in Applied Mathematics and Computational Science18(1), 1–28 (2023)

    work page 2023

  35. [35]

    SIAM Journal on Numeri- cal Analysis63(4), 1398–1426 (2025) https://doi.org/10.1137/23M1606812

    Biswas, A., Ketcheson, D.I., Roberts, S., Seibold, B., Shirokoff, D.: Explicit Runge–Kutta methods that alleviate order reduction. SIAM Journal on Numeri- cal Analysis63(4), 1398–1426 (2025) https://doi.org/10.1137/23M1606812

    work page doi:10.1137/23m1606812 2025

  36. [36]

    Published electronically at http://oeis.org (2025)

    OEIS Foundation Inc.: The On-Line Encyclopedia of Integer Sequences. Published electronically at http://oeis.org (2025)

    work page 2025

  37. [37]

    Hundsdorfer, W.H., Verwer, J.G.: Numerical Solution of Time-dependent Advection-diffusion-reaction Equations vol. 33. Springer, Berlin/Heidelberg (2003)

    work page 2003

  38. [38]

    BIT Numerical Mathematics22(2), 211–232 (1982) https://doi.org/10.1007/BF01944478

    Hairer, E., Bader, G., Lubich, C.: On the stability of semi-implicit methods for ordinary differential equations. BIT Numerical Mathematics22(2), 211–232 (1982) https://doi.org/10.1007/BF01944478

    work page doi:10.1007/bf01944478 1982

  39. [39]

    Hairer, C

    Hairer, E., Lubich, C., Wanner, G.: Geometric Numerical Integration, 2nd edn. Springer Series in Computational Mathematics, vol. 31. Springer, Dordrecht (2006). https://doi.org/10.1007/3-540-30666-8 25

    work page doi:10.1007/3-540-30666-8 2006