If a scaled local error estimator behaves as with approximately unchanged error scale, the displayed update targets a fixed local level. A safety factor aims below the acceptance boundary. For an embedded Runge-Kutta pair with order of a numerical method for the accepted output and for the companion formula, the difference generally has ; for the actual unscaled local truncation error of an order- formula, . The exponent belongs to the quantity being controlled. Dividing the defect by changes that exponent. This follows by solving for the next step.[ref]
Rejecting a trial means retrying from the same accepted time and state, with a smaller step. Advancing time or retaining the rejected output would change the initial value problem being approximated. Minimum and maximum change factors prevent extreme updates; a zero estimate requires finite capped growth. After acceptance, cap the next step at the remaining interval length. Detect a step that cannot advance time in floating-point arithmetic, and report failure rather than an endless retry. An adaptive step-size controller can use earlier estimates to smooth step sequences, but the memoryless rule here is sufficient to explain acceptance and rejection.[ref]
Consistency of a numerical method Created 2026-09-28 Updated 2026-10-14
For a one-step approximation to an ordinary differential equation, insert the exact solution and define the unscaled local truncation error
Consistency of a numerical method requires uniformly on the finite interval being approximated, under the stated smoothness assumptions. Merely asking is too weak: the useless update has defect , which tends to zero but generally fails the scaled test. Here denotes Big O notation. MIT uses this division-by-step convention.[ref]
For the Euler method, Taylor's theorem gives . A bounded second derivative therefore gives local truncation error and a scaled defect . Under a uniform Lipschitz continuity assumption on in , within a region containing the exact and numerical solutions, numerical stability controls accumulation over about steps. This yields global error with exact starting data. Consistency of a numerical method checks each step; numerical stability controls their combination.
For the five-point discretization of on the unit square, the displayed quadratic polynomial has exactly, nonnegative boundary values and . If and has zero boundary values, the discrete maximum principle applied to gives . The same estimate holds for , since . This turns a uniform local truncation error bound into a pointwise global error bound without inverting the full matrix.
Euler-Heun embedded pair 2026-10-14
The Euler-Heun 1(2) pair combines the explicit Euler method and Heun's method. Here 1(2) names their respective order of a numerical method; the accepted endpoint in this article is the second-order one. Its shared Runge-Kutta stages and endpoints are
The first output row in this Butcher tableau is Heun's method; the second is the explicit Euler method. Both outputs cost two evaluations of together, rather than one for Euler method plus two for a separately computed Heun's method.
For the initial value problem , , take . The first slope is ; the stage state is , giving . Therefore
The exact exponential function gives signed local defects approximately for the explicit Euler method and for Heun's method. Their difference is exactly . Thus the pair difference estimates the larger Euler defect; calling it the actual Heun defect would miss by a factor about fourteen here.
More generally, a step from on has
These original Taylor expansions separate three quantities: the accepted Heun's method has local truncation error and global error under suitable numerical stability, while the tested local error estimator is . For with held approximately fixed, use . A cube-root update would model a different quantity, the actual higher-order local defect.
Euler one-step defect 2026-10-31
For a smooth solution of , Taylor's theorem gives the displayed local truncation error when an Euler method step starts at an exact solution value. Thus the one-step defect is . Dividing it by gives a normalized local truncation error of ; these are different conventions for reporting the same residual. Also along the solution by the chain rule. A global error conclusion additionally requires the usual regularity and stability control over successive steps.
Explicit midpoint method 2026-10-31
This explicit numerical method is a two-stage Runge-Kutta method and a one-step method. For a sufficiently smooth ordinary differential equation, the chain rule gives , and Taylor's theorem shows that its local truncation error is . Uniform Lipschitz continuity of in its solution argument gives an increment-function Lipschitz constant at most . The Lipschitz one-step error bound then proves second-order numerical convergence on a fixed time interval. This method uses an intermediate prediction from ; the explicit midpoint linear multistep method instead requires a separate previous value .
Fourth-order correction of the nine-point Poisson stencil Created 2026-10-05 Updated 2026-10-14
For a smooth solution of , the corrected formula has unscaled residual and normalized local truncation error . The inverse interior matrix has Euclidean operator norm ; on grid points this gives unweighted Euclidean nodal error . Uniform derivatives up to the boundary are required for a grid-independent remainder bound.
Global discretization error Created 2026-10-05 Updated 2026-10-06
The global error is the difference between the computed and exact solutions after accumulated numerical steps, rather than the local truncation error of a single step started from exact data. An estimate for the stability of a numerical method controls how initial errors and local defects accumulate. For a stable order- time integrator with starting errors , local defects accumulated over steps give global error on a fixed time interval.
Heun's method 2026-10-14
Here Heun's method means the explicit endpoint-averaging Runge-Kutta method: take an Euler predictor , sample there, and average it with for the output. The second Runge-Kutta stage is evaluated at the predictor, not at the corrected endpoint. In particular it requires no implicit solve as the ODE trapezoidal rule does. Naming conventions vary between books; the formulas specify the method unambiguously.[ref]
For smooth , a Taylor expansion about gives , where is the state Jacobian matrix. Consequently the output is , matching the exact solution through degree two by the chain rule. This gives local truncation error and, with the usual bounded smoothness and Lipschitz continuity hypotheses on a fixed finite interval, global error . The Euler-Heun embedded pair compares this output with its Euler predictor.
Figure 1.
Euler-Heun stage geometry for y prime equals y: sample at the Euler predictor, then average slopes for the corrected endpoint
. Details.
For , and , the sampled slopes are and . Transporting the second slope to the starting point would end at ; the average slope instead ends at . The exact exponential function reaches . The magnified endpoint comparison separates the pair difference from the actual Heun local defect .
Suppose the increment function of a one-step method is Lipschitz continuous in its solution argument with constant , uniformly in time and step size. Let be the global error, and let be the exact-start local truncation error. The triangle inequality gives . Iterating and summing a geometric series proves the displayed estimate. Since , uniformly bounded and consistency error give global error on a fixed time interval when . Here denotes Big O notation. This is a constant-step form of the discrete Gronwall inequality.
Local error estimator 2026-10-14
A local error estimator approximates a numerical step's defect from the exact solution starting at the same input. Differences between formulas estimate error only after their asymptotic orders and constants are accounted for. Step-doubling error estimation, an embedded Runge-Kutta pair and predictor-corrector error estimation are common constructions. They do not by themselves bound accumulated global error.
Let be the exact endpoint of the ordinary differential equation started at the numerical value , and let be the candidate numerical endpoint. The signed local defect being estimated is . In contrast, actual global error compares with the solution of the original initial value problem, including all previous steps. The local truncation error convention inserts original exact data instead. These starting-value distinctions matter even when the same formula is used for every step. A practical local error estimator also relies on a smooth Taylor expansion; a small difference can reflect accidental cancellation rather than a rigorous upper bound.
Multiderivative multistep method Created 2026-10-05 Updated 2026-10-14
A multiderivative multistep method uses higher time derivatives as well as values and first derivatives from several time levels. For an autonomous ordinary differential equation , the second time derivative is by the chain rule. Its local truncation error must include the higher-derivative terms. Consequently the usual second-order barrier for A-stable linear multistep methods, which only use first derivatives, does not apply.
Normalized local truncation error Created 2026-10-14 Updated 2026-10-31
For a time update with one-step defect after exact solution values are inserted, the normalized defect divides by the time step . If , the normalized local truncation error is . A spatial mesh contributes its own powers: for a diffusion update, means . Under these are respectively second and fourth powers of . Neither normalization changes the actual order of a numerical method; it changes the power used to express its residual.
Order of a Runge-Kutta method Created 2026-09-24 Updated 2026-11-05
A Runge-Kutta method has classical order when its one-step map started from exact data agrees with the exact flow through powers for every sufficiently smooth ordinary differential equation. Its local truncation error is then , with a nonzero coefficient at that order for some problem if is the exact finite order.
For example, forward Euler method on , , returns , while the exact solution is . The first omitted term is quadratic, so the local error is and the method is first order. On a fixed time interval, Lipschitz continuity and a stable error-propagation estimate turn local defects into global error .[ref]
Matching the Dahlquist test equation checks only the linear stability expansion. The nonlinear Butcher order conditions must also hold; agreement on one scalar equation alone does not prove classical order. Stage accuracy is a separate property, so the number or individual precision of Runge-Kutta stages should not be substituted for the order of the final update.
Insert a smooth exact solution, so and . Expand about . The residual, left side minus right side, has the formal differential-operator symbol
Its odd symmetry makes every even coefficient vanish. The coefficients of are respectively
The next coefficient is , giving
Hence the local truncation error has exact defect order seven and the method has order six. The zero-step characteristic polynomial is , whose two unit roots are simple, so it is also zero-stable. With sufficiently accurate starting values and exact evaluations, the usual accumulation of defects gives sixth-order global error. This is the symmetric sixth-order two-derivative method; approximate evaluations of must preserve the required accuracy if that order is to be retained.
For an initial value problem, convergence of a numerical method means that on each fixed interval , the largest error at mesh points tends to zero as the step size tends to zero and the starting errors vanish. Order of a numerical method describes how rapidly an exact-start local defect decreases; stability of a numerical method controls propagation of that defect and of errors already present. These are related but different properties. A high-order expansion alone cannot ensure that accumulated error remains small.
For a one-step map , define the exact-start defect . Order means for sufficiently smooth solutions. The normalized local truncation error divides this defect by and is ; explicitly specifying the normalization prevents an apparent one-order discrepancy. Suppose the map satisfies the local Lipschitz bound
For the error recurrence is
Iteration, or the discrete Gronwall inequality, yields for . Stable propagation converts an order- local formula into global error , provided starting data are equally accurate. Consistency is the weaker requirement that the normalized local defect tend to zero.
For a linear multistep method, zero-stability concerns the recurrence at . Its roots must have modulus at most one, and unit-modulus roots must be simple. Roots outside the circle amplify errors exponentially in the number of steps; repeated unit roots introduce polynomial growth. The requirement is stronger than bounded propagation for one fixed number of steps: as on a fixed physical interval there are steps. Under the usual Lipschitz assumptions and consistent starting values, the Dahlquist equivalence theorem says consistency plus zero-stability is equivalent to convergence. Its sufficiency combines a bounded discrete recurrence propagator with the sum of local defects, while necessity follows by exciting an unstable homogeneous root or by observing the persistent error of an inconsistent formula.
For example, satisfies the first-order consistency identities, but its root polynomial is . For , a small starting error can excite a mode and destroy convergence as the number of steps increases. In contrast, simple parasitic unit roots such as the root minus one of a two-step centered method need not destroy zero-stability: they remain bounded, though they can create unwanted oscillations and affect practical behavior. Accurate starting procedures and suppression of accumulated parasitic error remain relevant.
Absolute stability asks a different question: for , does the discrete solution decay or stay bounded for a specified finite ? The Forward Euler method has and, for real , requires . It converges as , but a stiff decay mode can impose a severe practical step restriction unrelated to the accuracy needed for slowly varying components. The Backward Euler method has , hence is A-stable and L-stable. The trapezoidal rule has , is A-stable, but tends to minus one on a very stiff negative mode and can retain alternating transients rather than damping them rapidly.
Zero-stability therefore does not imply unrestricted absolute stability; the three-step family analyzed earlier in this paper is a particularly strong illustration. Conversely a stable large step need not be accurate. This distinction matters for a stiff equation. The Second Dahlquist barrier limits A-stable linear multistep methods to order at most two, whereas implicit Runge-Kutta methods can have much higher order and still be A-stable. That barrier concerns the specified method class, not every numerical time integrator.
Finally, the mathematical order depends on smoothness: nonsmooth solutions or forcing can lower observed rates. Roundoff, imperfect implicit solves and inaccurate derivative evaluations introduce additional defects. Smaller steps increase the number of such perturbations; truncation error alone does not describe arbitrarily fine computation. The useful organizing principle is consistency supplies vanishing defects, stability bounds their accumulation, convergence follows, and order quantifies its rate under the stated regularity and starting-error assumptions.
Write , . Every smooth exact solution of the advection equation here is . At , substitution into the scheme gives the unscaled residual
The constant, linear and quadratic Taylor coefficients vanish. Its cubic coefficient is
For a general smooth function the first-order residual is , so the proper normalized local truncation error on an exact solution is
Hence the scheme has formal order two for fixed general positive . At and , the displayed residual vanishes identically for every , not just to one higher Taylor order: the recurrence transports exactly sampled characteristics when both starting levels are compatible with the exact solution. This formal exactness does not imply stability. At there is no positive time step. The two-level recurrence also requires a starting procedure for ; one initial solution profile alone is not two independent starting levels.
Use the coefficient in the numerator, as verified in the PDF; the converted TeX lost its . Multiplying the method by gives the characteristic polynomials of a linear multistep method
For every admissible real , and , giving consistency of a numerical method. Factorization gives
The second root is . The root condition for a multistep method requires and forbids a repeated root at one. For real ,
At both roots equal one, violating zero-stability. For the second root lies strictly inside the unit disk. The Dahlquist equivalence theorem therefore gives
To find the order rather than only consistency, expand the exact-solution residual through the generating function:
Thus its unnormalized local truncation error begins with when , and the global method order is exactly two for every convergent parameter. For comparison, has formal order three but is not convergent because of its double unit root:
These orders assume a smooth solution, the usual local Lipschitz condition and starting approximations of the required accuracy.
Consider on a rectangle, with prescribed Dirichlet boundary conditions, and assume initially that the solution is smooth up to the boundary. A finite difference method chooses grid values as unknowns, approximates the differential operator by a local stencil, incorporates the boundary data, and solves the resulting sparse linear system. Accuracy of a stencil by itself is insufficient: one also needs a well-posed discrete problem, a stability estimate uniform under refinement, and a sufficiently accurate algebraic solve.
For a square grid of spacing , centered second differences give the five-point Laplacian
Expanding each opposite pair in a Taylor series gives
Thus the standard equation has second-order normalized local truncation error. Known boundary values move to the right-hand side; they are not iterated as independent unknowns. Nonrectangular geometry requires a consistent boundary approximation rather than silently retaining a full interior stencil outside the domain.
The negative interior operator has positive diagonal and nonpositive off-diagonal entries. For a grid vector extended by zero on the boundary,
The edge sum includes links to the fixed boundary. Connectivity makes the matrix positive definite, proving uniqueness and providing an energy structure. The same sign pattern gives a discrete maximum principle: at a negative interior minimum, is nonpositive. If and the boundary is nonnegative, equality can only persist by propagating that minimum through all neighbors to the boundary, a contradiction. Therefore .
On the unit square, the barrier satisfies exactly. If has zero boundary values and , apply the maximum principle to to obtain
The bound is independent of , so the consistency error gives an nodal error. It also separates the algebraic residual from the discretization error: solving the linear system much less accurately than would spoil the spatial accuracy.
A direct higher-order design is possible. In one coordinate,
has fourth-order accuracy. Adding the corresponding formula in the other coordinate gives a wider stencil. Its advantages must be balanced against boundary closures, more distant couplings, and the possible loss of monotonicity. Higher order does not automatically imply a stable or geometrically convenient scheme.
The Mehrstellen method takes a different route: it uses the Poisson equation to eliminate leading truncation derivatives while keeping a compact unknown stencil. Define the nine-point finite-difference stencil
Here each sum has four terms. Equivalently . Expanding gives
The uncorrected nine-point operator is only second order for a general function. But implies . Moving this known leading defect to the source and approximating its Laplacian with gives
Indeed , and multiplication by makes that replacement's error fourth order. In explicit compact form the right-hand side is . Only known source samples, not additional distant solution unknowns, enter this correction. The corresponding unscaled stencil residual is . If , the leading term already vanishes, which explains harmonic superconvergence of the nine-point stencil.
For zero boundary data, has axial link weights and diagonal link weights . Their energy sum proves positive definiteness, and their sign pattern again proves the discrete maximum principle. The same quadratic barrier is exact for , so
For smooth solutions and accurately imposed boundary values this proves fourth-order nodal convergence of the corrected scheme. Singular corners, rough sources or low-order boundary treatment can reduce the observed rate; a formal interior expansion is not a global regularity theorem.
More generally, compact high-order schemes choose a solution stencil and a source stencil together, match their Taylor expansions, and use differentiated forms of the PDE to remove derivatives that would otherwise need a wide solution stencil. Variable coefficients require their derivatives and mixed terms to be treated consistently; the constant-coefficient identity cannot be reused unchanged for a different elliptic operator. On irregular meshes, conservation and consistent fluxes may be more useful starting points than a symmetric Cartesian formula.
Finally, discretization and solution cost are coupled. The five-point Dirichlet matrix has eigenvalues proportional to sums of squared sine frequencies, so its smallest eigenvalue stays of order one while its largest is . Its condition number grows like ; simple relaxation therefore becomes slow. Sparse direct elimination, conjugate gradient method with appropriate matrix preconditioning, and the multigrid method are alternatives. A sound design combines local accuracy, boundary fidelity, mesh-independent stability, and an algebraic solver whose residual is small enough for the intended error tolerance.
Write , , and assume are sufficiently smooth for the following Taylor expansions. The symmetric half-grid diffusion consistency calculation gives
where
The normalized local truncation error of the full update is consequently
The PDF prints an advection equation, , whereas this stencil approximates conservative diffusion, . For the literal printed equation the leading defect is , which generally does not vanish: the method is inconsistent, with an normalized defect.
For the intended variable-coefficient conservative diffusion equation, the first two terms cancel and . Under the parabolic step scaling , this is , while the unnormalized one-step local truncation error is . These two orders use different normalizations. With numerical stability and compatible smooth initial and boundary conditions, the intended diffusion problem has global error on a fixed time interval. Positivity and boundedness of alone do not justify the displayed Taylor expansion; coefficient regularity is also needed for that error formula.