Let , and . Along a smooth solution, . The exact difference contains only odd derivatives. Matching terms through degree four therefore requires
The first two conditions give , which also implies . With , the remaining equations are and . Hence
The fifth spatial moment is . After dividing the exact-minus-scheme residual by , its first possible nonzero term is
This is the fourth-order two-step advection stencil. Thus the normalized local truncation error is at fixed positive Courant number, as required. At or the formula translates the physical advection branch exactly, but consistency of a numerical method alone does not decide stability of all two-level perturbations.
Choose a partition , with the usual nodal piecewise-linear hat functions , including the half hats at the endpoints. Write with . Differentiating the Ritz method energy with respect to , , gives
On an element of length , its two basis functions are and . Direct integration yields the element matrix
For example the first weighted diagonal is ; the off-diagonal is . This gives the affine-weighted hat mass matrix and explicit equations on any partition by assembling adjacent elements.
For the uniform choice , , the assembled interior equations are
The last node contributes only one element, so its equation is
In the first interior equation the known term moves to the right, giving . For only the endpoint equation is needed. The last row enforces the weak Neumann boundary condition, so no artificial value outside the interval is introduced. The matrix on is a symmetric positive-definite matrix: its quadratic form is for every nonzero trial variation . The discrete minimizer is therefore unique.
Take . The Airy equation is . To discuss positivity of this linear operator, use its homogeneous variation domain . Integration by parts gives
The boundary term is zero at both ends. The same calculation with two different functions proves that the associated form is a symmetric bilinear form. Positivity here concerns the homogeneous domain; the condition instead defines an affine set of admissible solutions.
For the variational problem define
and minimize
The Sobolev trace at zero is meaningful in . No derivative boundary value is imposed on this trial space: the Neumann boundary condition will emerge naturally. For , Cauchy-Schwarz inequality and integration give
so . Thus is a bounded, symmetric, coercive bilinear form on the Hilbert space . The Lax-Milgram theorem says that a bounded coercive bilinear form and a bounded linear functional determine a unique weak solution. Apply it to
and set . This is precisely the first variation condition for .
Compactly supported test functions first yield in the sense of distributional derivatives. Since , the solution is in . Integration by parts then leaves for every , hence . The other boundary condition is built into . Conversely a solution of this boundary value problem satisfies the first variation condition. Finally,
This proves existence, uniqueness and the global minimum characterization, including the natural boundary condition. It is the mixed-boundary Airy energy principle.
For , an exterior polynomial root of persists by continuity for sufficiently small negative real in the stability polynomial . Hence these parameters cannot be A-stable.
For , the principal polynomial root at is simple. Let denote its analytic logarithm near zero. Using the expansion in part (b), and , solve the polynomial root equation to obtain
Indeed the residual at is , and the correction in cancels it by multiplication with . Choose . This lies strictly in the left half-plane, but
for small positive . Thus , proving linear instability within the A-stability domain. This is a direct principal-root obstruction to A-stability, rather than an appeal to an order-barrier theorem alone.
Finally, at , the full stability polynomial is
It retains a double unit polynomial root for every negative , allowing growing parasitic solutions. Therefore
The canceled Backward Euler method recurrence at is A-stable, but its admissible solutions and starting relations differ from those of the original method.
Apply the exponential-symbol order criterion for a multistep method, which follows by substituting the bilateral shift operator into the exact-solution residual. Direct expansion gives
All coefficients through degree three vanish. If , the fourth-degree coefficient is nonzero, so the normalized local truncation error has exact order three. At , the fourth-degree coefficient vanishes and the fifth-degree coefficient is , which is nonzero. Thus
This is the formal order of the full recurrence. The order-four member is not zero-stable, so it is not convergent. All the convergent members from part (a) have order three. At the degenerate value , the full recurrence still has formal order three, while cancellation of its common factor gives Backward Euler method of order one; that cancellation changes the method rather than its order calculation.
The first characteristic polynomial factors as
Also and , so the order conditions for a linear multistep method hold for every . Convergence of the full linear multistep method recurrence requires, in addition, the root condition for a multistep method: all polynomial roots of lie in the closed unit disk and its polynomial roots on the unit circle are simple.
For , write with . The other polynomial roots are , distinct from each other and from 1, so the root condition for a multistep method holds. For , the two quadratic polynomial roots are real and reciprocal, and one has modulus greater than one. At , ; at , . Both endpoints violate simplicity. Consistency of a numerical method plus zero-stability therefore gives
The conclusion applies to the recurrence as printed, including its starting data. Canceling common factors is not innocuous: at the polynomials contain , and at they contain . The reduced recurrences exclude parasitic solutions of the original one. This is the common-factor cancellation defect in a multistep recurrence, so those parameter values are not added to the interval.
Use a spatial discrete Fourier mode . Its recurrence and amplification polynomial of a multilevel finite difference scheme are
The polynomial roots are . For both have modulus one and are separated. Since , a fixed bounds their separation below by . The two-level companion matrix is therefore diagonalizable with uniformly bounded eigenvector matrix and inverse. To see the uniform bound explicitly, write . Its determinant is ; all its entries have modulus one, and those of are bounded by the reciprocal of the root separation. The companion matrix powers are , bounded independently of frequency and step number. Parseval's identity converts this into a mesh-independent discrete bound for arbitrary perturbations at the two initial levels.
If , choose . Then , and one polynomial root has modulus . On periodic meshes with the number of points divisible by four, this is an actual grid mode, producing exponential linear instability.
At the same phases give . The companion matrix is not a scalar matrix, so the double polynomial root has a nontrivial Jordan block. The solution has bounded starting amplitudes but grows like . On a fixed physical time interval is of order , so no mesh-independent stability bound exists. Consequently
is the stability interval under the standard two-level definition. Testing only polynomial root moduli would misleadingly include the endpoint. A special startup selecting the non-growing branch at that endpoint restricts the perturbations and does not establish stability of the full recurrence. This is the two-dimensional leapfrog stability threshold.
Put , , with fixed positive . Insert a smooth exact solution and divide the residual by :
Taylor expansion gives
The differential equation cancels the leading part, leaving normalized local truncation error . Since contains mixed derivatives, there is no fixed positive Courant number canceling this leading error for every smooth solution. Thus the method is second order in time and space. Global second-order convergence also requires stability and initial values at both time levels accurate enough to retain this order.
Use the Gauss collocation coefficient construction. Find the polynomial roots of the shifted Legendre polynomial in , and form
Then set
These are explicit polynomial integrals. For example, if , calculate and . The resulting implicit Gauss collocation method has order ; its Gaussian quadrature nodes integrate polynomials through degree exactly. The one-stage example is the implicit midpoint rule, with and . No search over nonlinear Runge-Kutta order conditions systems is needed.
Write the Runge-Kutta method coefficients as , , and let be the all-ones column. On the test equation , with , the stage vector satisfies and the next value is . Hence its stability function is
Both polynomials have degree at most , and the denominator has value one at zero. Order of a Runge-Kutta method implies that this test-equation approximation agrees with through degree :
By uniqueness of the diagonal Padé approximant, . Explicitly,
The permitted diagonal Padé approximant property says this rational function has no poles in the closed left half-plane and satisfies there. Therefore
There is also no hidden stage-solve singularity from a canceled denominator: the Padé approximant numerator and denominator are coprime and each has degree . The original determinant denominator already has degree at most , so, with its normalization at zero, it equals and cannot contain an additional canceled factor. This proves maximal-order Runge-Kutta methods are A-stable.
The order- divided difference defining is a finite linear combination of values at its knots. Hence integration can be interchanged with it directly. For a knot value ,
It follows that
The final equality holds because the order- divided difference of a degree- monic polynomial is its leading coefficient, one. Thus
This proves the unit-integral normalization of a B-spline. Its support is contained in : below all its knots, the sampled truncated power agrees with a degree- polynomial, annihilated by the order- divided difference; above all its knots, every sample is zero. No mass is omitted by integration over .
Since , the corresponding partition-normalized B-spline has integral . Its normalization is different from that of .
For this part relabel one consecutive block of distinct knots as . This is a local reindexing; no extra global knot is being introduced. Let . Its Lagrange interpolation polynomial through the knot values is
Each quotient is monic of degree . Comparing leading coefficients gives
The same leading coefficient is the order- divided difference , by the Newton interpolation polynomial representation. Multiplying by therefore gives the explicit B-spline formula
Distinct knots make every denominator nonzero. For the degree-zero truncated power is interpreted as an interval-indicator convention; values at individual endpoints do not affect its integral.
Apply the general B-spline interpolation operator norm estimate in part (I) to the inverse norm from (b):
and hence in particular , independently of . The strict finite-dimensional estimate uses the actual finite collocation matrix; it does not require periodic closure, natural boundary conditions or endpoint knot repetitions.
There is also a direct shorter verification of the uniform bound. For an index maximizing , the triangle inequality gives
with an even larger margin when a neighbor is missing. This is the inverse infinity-norm bound from diagonal dominance, and again gives .
The B-spline collocation matrix in the cubic case is the finite tridiagonal matrix
For , this is simply the one-entry matrix .
To evaluate the inverse norm exactly, let . Then , where has ones immediately above and below the diagonal and zeros elsewhere. Since , the convergent Neumann series
has nonnegative entries and proves invertibility. Since , its absolute entries equal those of . The absolute row sums are therefore the entries of .
The equation becomes
Its constant particular solution is ; the two characteristic roots of the homogeneous recurrence are and . Imposing the two boundary values gives
This can also be checked by direct substitution, since . The numerator is smallest at the central index or central pair of indices: its successive difference is . Thus, with , the exact inverse norm of midpoint cubic spline collocation is
Equivalently, if the fraction is , and if it is . The finite norm is not exactly three: it equals for , two for , and tends to three as grows. Hence three is the sharp dimension-independent bound for these inverse matrix norms.
Let be the order- Cardinal B-spline supported on . The Cox-de Boor recurrence on unit-spaced knots reads
Start with . The order-two hat satisfies and vanishes at its support endpoints. Substituting these values into the recurrence gives and then
The continuous cubic spline is zero at and outside its support. Since and , this yields the midpoint cubic spline collocation values
Only indices are retained. In particular no nonexistent spline is added to an endpoint row.
Since on and the coefficients are summable, the Weierstrass M-test gives uniform convergence of this positive lacunary Chebyshev series and makes its sum continuous.
Let be the largest integer with , and set when . The candidate and its tail size are
The empty sum for is zero. The candidate has degree at most . The triangle inequality bounds its error by the proposed tail sum, and equality holds at , since every Chebyshev polynomial has and all tail coefficients are positive. It remains to prove that another polynomial cannot improve this error.
Put , the first omitted degree, so . At the points for , every omitted term has
because is odd. Uniform convergence therefore gives
These distinct points run in decreasing order; reversing their order still gives alternating signs. Since , select consecutive ones and apply the Chebyshev alternation theorem. It proves both optimality and uniqueness of the displayed candidate. Thus the best polynomial and the error stay unchanged between successive powers of three; this exact truncation property depends on the synchronized tail signs, not on a general rule that truncating a Chebyshev expansion is optimal.
For real-valued , there is a unique best uniform approximation , where means degree at most . If , the Chebyshev alternation theorem characterizes it by the existence of ordered points
and a sign such that
Conversely, a polynomial in whose error has this alternation is the best one. If the best error is zero, itself is the exact polynomial approximant. The real-valued qualification matters: this sign-alternation formulation is not the corresponding criterion for arbitrary complex-valued approximation.
The obstruction to a strict improvement is concrete. If had smaller error at every alternation point, would have alternating strict signs there. The intermediate value theorem would give at least distinct zeros, impossible for a nonzero polynomial of degree at most .
Take any trigonometric polynomial . The reproduction property in (a) gives
By (b),
Taking the infimum over yields the polynomial reproduction error bound for this operator. With and , it becomes
This is a near-best uniform approximation estimate; it follows from reproduction and boundedness without requiring the de la Vallée Poussin sum itself to be a best approximant.
The difference between the two cumulative sums of Fourier partial sums is
Hence the de la Vallée Poussin sum satisfies
Using the uniform-norm contraction of Fejér summation on each term and the triangle inequality in the supremum norm,
This proves the requested operator norm bound. When the identity reduces to and the bound is one.
A trigonometric polynomial of degree at most has zero Fourier coefficients outside . Its Fourier partial sum therefore equals for every . Averaging the sums with gives
Thus the de la Vallée Poussin sum reproduces the entire degree-at-most- approximation space, including when the actual degree is smaller.

Pinned article: Introduction to the OurBigBook Project

Welcome to the OurBigBook Project! Our goal is to create the perfect publishing platform for STEM subjects, and get university-level students to write the best free STEM tutorials ever.
Everyone is welcome to create an account and play with the site: ourbigbook.com/go/register. We belive that students themselves can write amazing tutorials, but teachers are welcome too. You can write about anything you want, it doesn't have to be STEM or even educational. Silly test content is very welcome and you won't be penalized in any way. Just keep it legal!
We have two killer features:
  1. topics: topics group articles by different users with the same title, e.g. here is the topic for the "Fundamental Theorem of Calculus" ourbigbook.com/go/topic/fundamental-theorem-of-calculus
    Articles of different users are sorted by upvote within each article page. This feature is a bit like:
    • a Wikipedia where each user can have their own version of each article
    • a Q&A website like Stack Overflow, where multiple people can give their views on a given topic, and the best ones are sorted by upvote. Except you don't need to wait for someone to ask first, and any topic goes, no matter how narrow or broad
    This feature makes it possible for readers to find better explanations of any topic created by other writers. And it allows writers to create an explanation in a place that readers might actually find it.
    Figure 1.
    Screenshot of the "Derivative" topic page
    . View it live at: ourbigbook.com/go/topic/derivative
  2. local editing: you can store all your personal knowledge base content locally in a plaintext markup format that can be edited locally and published either:
    This way you can be sure that even if OurBigBook.com were to go down one day (which we have no plans to do as it is quite cheap to host!), your content will still be perfectly readable as a static site.
    Figure 5. . You can also edit articles on the Web editor without installing anything locally.
    Video 3.
    Edit locally and publish demo
    . Source. This shows editing OurBigBook Markup and publishing it using the Visual Studio Code extension.
  3. https://raw.githubusercontent.com/ourbigbook/ourbigbook-media/master/feature/x/hilbert-space-arrow.png
  4. Infinitely deep tables of contents:
    Figure 6.
    Dynamic article tree with infinitely deep table of contents
    .
    Descendant pages can also show up as toplevel e.g.: ourbigbook.com/cirosantilli/chordate-subclade
All our software is open source and hosted at: github.com/ourbigbook/ourbigbook
Further documentation can be found at: docs.ourbigbook.com
Feel free to reach our to us for any help or suggestions: docs.ourbigbook.com/#contact