Use real-valued continuous functions and real trigonometric polynomials. The trigonometric Chebyshev alternation theorem says that is the unique best uniform approximation to if and only if there are distinct points, ordered within one period,
Thus the required number of alternating extrema is . When the error is zero, the assertion is interpreted trivially and is unique. The count reflects and the fact that a nonzero degree- trigonometric polynomial has at most zeros in one period. In particular a strictly better candidate would force its difference from to have alternating signs at these points and hence at least zeros, counting the closing interval around the circle. The trigonometric Chebyshev alternation theorem also gives necessity and uniqueness.
First prove necessity. If polynomials with integer coefficients converge to in the supremum norm, their values at and are convergent sequences of integers. Each such sequence is eventually constant, since a convergent sequence is Cauchy and two distinct integers are at distance at least one. Its limit is an integer. Thus .
For sufficiency we also need uniform convergence of the Bernstein polynomial to . Here is a direct argument. For , write . The binomial distribution has and . For every , uniform continuity of and the Chebyshev inequality give
First make small and then large; the bound is uniform in . Hence .
If the endpoints are integers, part (b) gives . Each coefficient multiplying in is an integer. Expanding by the binomial theorem therefore gives integer coefficients in the ordinary monomial basis. The triangle inequality proves . Thus the integer endpoint condition is necessary and sufficient for integer-coefficient polynomial approximation.
The integer endpoint hypothesis is important. Let . The floor function gives , while because and are integers. Therefore the rounded Bernstein polynomial satisfies
For , the successive terms have ratio , so the geometric series estimate gives
The last inequality follows by maximizing at . The same bound holds on by reflection. On , each factor and is at most , giving . Thus
Here ; for the difference vanishes. Without the integer endpoint hypothesis the claim would fail: at the error is , independently of .
The Bernstein polynomial is the positive weighted average
For , the Bernstein basis weights are nonnegative and the binomial theorem gives . Hence . In particular,
This is the contraction property of the positive linear operator on continuous functions defined by the Bernstein polynomial.
Fix the standard Jacobian matrix convention . With column gradients, the commuting velocity derivative for kinetic transport is
If the symbol is instead defined with derivative indices as rows, the printed expression is this same formula. Under the standard convention it needs a transpose.
To prove the claim, put . Differentiating in gives
The terms cancel, so each component of satisfies the transport equation. Multiplication by the smooth coefficient and differentiation preserve the same uniform compact support, hence componentwise. At , . The L2 norm calculation in part (d), summed over components, gives
One can also differentiate the explicit solution: the term cancels the derivative of , leaving the initial velocity derivative transported along the characteristic curves.
The transpose matters. For the shear , but . The printed field satisfies , which is nonzero for a generic compactly supported datum. The untransposed formula is false with the standard convention.
The displayed estimate has two independent defects. Even for , its right side cannot use . At , choose nonnegative smooth functions with compact support and prescribe . This is a legitimate transported solution with . The proposed inequality would require
Velocity dilation makes arbitrarily large, while the other factors stay fixed. Thus no universal works with the same-time mixed Lebesgue norm.
There is a second issue after replacing by : the Jacobian determinant bounds give a local diffeomorphism, not necessarily a one-to-one map. Suppose additionally that is injective. With , the change of variables formula and give
Surjectivity is not needed because the domain of integration can be enlarged. With injectivity, the corrected estimate has . The upper bound is unnecessary.
More generally, dispersion with a nonlinear velocity map uses the area formula to sum over inverse branches. Define the weighted inverse multiplicity
If , the same argument gives the corrected estimate with . At most inverse branches give . Thus the missing global assumption concerns multiplicity, not merely local volume distortion.
For a noninjective map with constant Jacobian determinant giving a counterexample in two dimensions, write and set
The polar coordinates calculation gives , yet for every integer . To turn this into a failure of the estimate, take a small open disk about that avoids the origin. For each of inverse branches over , choose a smooth cutoff supported on that branch, where has support strictly inside and is extended by zero. These velocity supports are disjoint. Set
Then , independent of , whereas at a fixed and suitable ,
The last integral is positive and independent of . No finite exists for this fixed , although . In one dimension, by contrast, a nonvanishing derivative has a constant sign, so the lower derivative bound makes a global diffeomorphism.
For the free transport equation, the mixed Lebesgue norm dispersion estimate is
Indeed, use the method of characteristics and then the change of variables :
Taking the essential supremum over proves the result. This decay measures the spreading of a velocity average; the total phase space Lp norm is conserved rather than decaying. The datum on the right is the initial datum, not the solution at the same time.
For two solutions with the same initial datum, let . Their uniform compact support on puts and in one compact region of phase space. Smoothness there implies that is finite and continuously differentiable, by differentiation under the integral sign. The transport equation and integration by parts in yield
There is no boundary contribution because of compact support, and . Since , almost everywhere and then everywhere by continuity. The solution is unique in . The same calculation gives Lp conservation for incompressible transport at .
Choose with and let , which is finite by continuity. Every point transported from satisfies
Hence any
contains the support of a function in the required phase space ball throughout . Equivalently, the image of the compact set under the characteristic flow map is compact. This uniform support control puts the solution in and justifies the following differentiations under the integral.
The chain rule gives along each characteristic curve. Pulling the initial datum back by the inverse characteristic flow map therefore gives
Direct differentiation also gives , so this is a classical solution of the transport equation. Composition with a smooth function makes smooth jointly in ; nonnegative initial data remain nonnegative.
If , then . A continuous image of a compact set is compact. Thus compact support is preserved at every finite time; there is no claim of one fixed compact support for all .
Use to label the initial point in phase space. The characteristic equations for a transport equation are and , with and . Since the velocity remains fixed, the ordinary differential equations integrate directly:
This characteristic flow map exists for all real times, even if grows rapidly at infinity: each trajectory samples just one fixed velocity. Its inverse is , and its Jacobian matrix is the block triangular matrix
Consequently its Jacobian determinant is one. The nonzero Jacobian determinant of is unnecessary for this construction.
The Dirichlet difference Laplacian has an orthogonal sine basis with eigenvalues , where for . On a mode the theta method amplification is
If , the denominator is positive and
Thus for every mode and every . Orthogonality gives a uniform discrete L2 norm contraction, not merely individual eigenvalue estimates.
If , a positive makes the implicit matrix singular. If , becomes arbitrarily large. If , tends to , whose magnitude is greater than one. Hence unconditional stability holds exactly for .
For completeness, is contractive under ; the mesh-independent sufficient restriction is . This follows from the same squared-modulus difference and keeps the denominator positive even for negative . At , very high frequencies approach amplification , so A-stability does not imply damping of rough heat data. At the Backward Euler method damps them strongly; this is the rough-data damping distinction for the theta method.
Take the endpoint values and evolve the interior indices . This matches on . If the printed upper index is retained, the odd ghost extension makes its boundary equation identically compatible; it does not add an interior unknown.
Writing and , this is the theta method with parameter :
For a smooth exact heat solution, its normalized residual at is
Thus the generic consistency orders are first in time and second in space for , and second in both time and space for , the Crank-Nicolson method. These are independent-step orders. Along the special parabolic refinement , the leading combined residual cancels if , giving a higher combined smooth-data order; it does not change the generic temporal order at independently varied meshes.
The given L2 initial condition need not have pointwise values or four derivatives. Initialize by a bounded cell-average or projection operator, as in L2-compatible initialization of grid data. The Taylor estimate applies to smooth data or sufficiently positive times after heat smoothing, rather than an unqualified uniform high-order estimate at the initial instant for arbitrary rough data. Stability and convergence for rough data are distinct from this consistency order.
Every convergent member is also A-stable. For , put and calculate its boundary-locus test for multistep A-stability:
The denominator never vanishes on the unit circle in this parameter interval. It follows that
Therefore no amplification root can cross the unit circle as moves in the open left half-plane. For small negative real , the root near one moves to , inside the circle; the other begins at , also strictly inside. The leading coefficient never vanishes in the left half-plane. Continuity of polynomial roots and connectedness of that half-plane now keep both roots inside everywhere. On its imaginary boundary the same locus has zero real part only at , corresponding to ; that root is simple because .
At , the two interlaced subsequences follow the trapezoidal rule with step :
The right side has modulus at most one for . Its unit-circle roots are simple; the double zero at lies strictly inside and is harmless. Thus the endpoint is also A-stable. Combining this with part 1, the requested range is , including the BDF2 method at .
The characteristic polynomials of a linear multistep method are
Expanding the exponential-symbol order criterion for a multistep method gives
Thus the formal local consistency order is two for , and three at , since the fourth-degree coefficient there is . This exceptional formal order is degenerate: , so it is not a convergent third-order time integrator.
For zero-stability, factor . The root condition for a multistep method requires every root to have modulus at most one and every unit-modulus root to be simple. For real it holds exactly for
At zero the second root is , simple and admissible; at two the root one is double and inadmissible. Outside the closed interval the second root has modulus greater than one. The Dahlquist equivalence theorem says that a consistent linear multistep method is convergent exactly when it is zero-stable, for a locally Lipschitz initial-value problem with consistent starting values and solvable implicit steps. Hence the displayed interval is also exactly the convergence range, and every convergent member has order two.
The common-factor cancellation defect in a multistep recurrence explains why the exception cannot be repaired just by canceling . For , its recurrence allows linear growth from a vanishing but insufficiently small starting perturbation: , gives , diverging at a fixed positive time. Canceling the factor silently imposes an additional starting relation and changes the method.
Use the discrete norm , with periodic indices. Two independent arguments prove uniform stability of a numerical method.
For Fourier stability analysis, a discrete Fourier mode with has semidiscrete eigenvalue
Its real part is nonpositive, so . The unitary discrete Fourier transform and discrete Parseval identity then give for every , and fixed real .
For the energy method, denote the centered first difference by and the forward difference by . Periodic discrete summation by parts makes skew-adjoint and gives . Therefore
The constant mode is preserved rather than strictly damped. Both proofs apply equally to differences of two solutions. The semidiscretization is contractive in the discrete L2 norm. A later choice of a time integrator would require its own stability analysis; no time-step condition is needed for this continuous-time result.
Let and sample a sufficiently smooth periodic exact solution. The central finite differences have expansions
and
The semidiscrete right side therefore equals . The spatial consistency of a numerical method is second order, for each fixed . There is no time discretization in the stated ordinary differential system, so no separate temporal approximation order is being assigned. The Taylor statement concerns smooth solutions; L2-compatible initialization of grid data handles nonsmooth initial data without undefined point sampling.
For the Dahlquist test equation with , the stage equations give the stability function . Substitution simplifies it to
The denominator zeros are , so no pole lies in the closed left half-plane. Moreover,
Thus throughout that half-plane: the method is A-stable. On the imaginary axis its amplification modulus is one. Since as , it is not L-stable; linear stability does not guarantee strong damping of very stiff modes.
The coefficients are those of the three-stage Lobatto IIIA method. Set , and . Direct calculation gives all eight fourth-order conditions for a Runge-Kutta method:
Powers of are componentwise. These Butcher order conditions prove order at least four for a smooth general nonlinear vector field; matching a scalar test equation alone would not suffice. A necessary fifth-order condition fails:
Hence the method has order exactly four, with one-step local truncation error and, under stability and smoothness hypotheses, global error .

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