For a finite-dimensional integral, . Expanding a polynomial interaction formally and interchanging each finite-order derivative with the regulated integral gives
where is the free Gaussian integral. This is a perturbation-series identity; it does not assert convergence of the full expansion for an arbitrary potential.
For the Minkowski functional integral, each field insertion is generated by
Expand , replace its field factors by these functional derivatives, and take the derivatives outside the free integral. The result is polynomial interactions as source differential operators:
If the field interaction term contains derivatives of the field, the corresponding spacetime derivatives act on the source-insertion operators as well. For the scalar polynomial interactions the operator in the exponent is
The free quadratic source exponential then supplies all Wick contractions, reproducing the Feynman rules. For normalized correlation functions, divide the resulting interacting functional by ; the interaction operator changes this denominator too, so it cannot simply be ignored.
Use the same metric signature as above and define the Minkowski generating functional by
Here . Integration by parts writes the quadratic action as with ; the infinitesimal imaginary part damps the oscillatory Gaussian functional integral and selects the vacuum Feynman propagator. Abbreviating repeated spacetime integrations, completing the square gives
Translation of the regulated integration variables therefore yields
The source-independent normalization contains the regulated functional determinant and its oscillatory phase; it cancels in . Define . The Gaussian evaluation of a free scalar generating functional is equivalently .
With Fourier convention , acts as . The inverse-kernel equation then gives
This is the time-ordered free two-point correlation function: differentiating twice with respect to and multiplying by recovers . It obeys in the distributional limit. The is essential: the algebraic denominator alone would not distinguish Feynman, advanced and retarded Green functions.
A linear operator on finite-dimensional vectors acts by . For a field the discrete index becomes a continuous coordinate, and a suitable operator is represented by an integral kernel, possibly a distribution:
The identity operator has kernel . If an inverse exists on the chosen domain and with the chosen boundary prescription, its kernel satisfies
with the analogous reversed identity. A differential operator is a distributional kernel: for example, . Its inverse is a Green function. Boundary conditions, a causal prescription and the treatment of zero modes are part of the inverse's definition, not consequences of formal matrix notation. An operator with an untreated zero mode does not have an inverse on the full field space.
The functional derivative is defined through the first variation: for arbitrary smooth test variation ,
In particular,
For a symmetric kernel, differentiating gives . The two terms from varying the two fields coincide; for a nonsymmetric kernel the symmetric part appears instead. Integration by parts is understood when is differential and the boundary variations vanish. A finite-volume, finite-mode regulator makes these statements ordinary matrix identities before a continuum limit.
Because is a real symmetric positive-definite matrix, its inverse exists and all its eigenvalues are positive. Completing the square gives
The translation has unit Jacobian determinant. Choose an orthogonal matrix with , where . The change also has absolute Jacobian one. The remaining integral is the product of one-dimensional Gaussian integrals, each equal to . Consequently the linear-source multivariate Gaussian integral is
The determinant square root is the positive one, as required by convergence of the original real integral.
The relevant boundary here is the upper critical dimension: above it the critical interactions become negligible at long distances and the ordinary mean-field critical exponents apply; at it there can be logarithmic corrections. This differs from the lower critical dimension, below which a particular kind of ordering may fail to exist.
Normalize the coefficient of in the Landau-Ginzburg theory. Under a change of length scale , its invariance gives the engineering dimension . The quartic interaction then scales as
It is a relevant operator below four dimensions and an irrelevant operator above four. Thus the ordinary scalar transition has upper critical dimension . At a tricritical point, where the quartic interaction is tuned away, the sextic coupling scales as , giving the tricritical upper critical dimension . This is the upper critical dimension of an even scalar interaction applied to powers four and six.
The Ginzburg criterion makes the reason for failure of the Landau approximation quantitative. In the ordered quartic theory, , and . The fluctuation of the order parameter averaged over a correlation volume has the scale
Compared with the squared mean order parameter, the Ginzburg criterion therefore gives
For this ratio grows without bound as the correlation length diverges. The assumption that fluctuations are small relative to the ordered value fails arbitrarily near the transition, and the interacting renormalization-group fixed point determines the critical exponents. For the ratio tends to zero. At scale counting is marginal, and the running quartic coupling produces logarithms.
Similarly, at the tricritical point , so the multicritical Ginzburg ratio scales as . The Landau-Ginzburg theory remains a useful effective description below its upper critical dimension; it is its fluctuation-free exponent calculation that fails, not the possibility of including fluctuations in the theory.
A tricritical point is the meeting of a line of continuous phase transitions and a line of first-order phase transitions. In the symmetric scalar Landau free energy, it occurs at with . Two independent controls are needed to tune both coefficients. The continuous transition lies at , and the tricritical three-phase line found above has ; both approach the tricritical point. At zero conjugate field the three coexisting states on this first-order line are the disordered state and the two opposite ordered states.
Setting makes the leading stabilizing interaction sextic. The nonzero minimum obeys , so the order-parameter critical exponent along this tuned direction is . Substitution gives
and two temperature derivatives give the heat-capacity critical exponent . At , the equation of state is , giving the critical-isotherm exponent .
For magnetic susceptibility, differentiating gives . This is above the transition and below, hence . The quadratic fluctuation kernel still has a mass proportional to , giving and . Thus the mean-field tricritical exponent calculation yields
Detuning gives tricritical crossover scaling; one must specify the approach to the tricritical point when quoting these critical exponents. The sextic interaction has upper critical dimension three, so sufficiently strong long-distance fluctuations change these mean-field predictions below three dimensions.
Critical exponents describe leading singular powers near a continuous phase transition. Using the reduced temperature , the standard definitions include
These are the order-parameter critical exponent, magnetic-susceptibility critical exponent, heat-capacity critical exponent and correlation-length critical exponent. The critical-isotherm exponent is defined by . The anomalous dimension gives the critical long-distance connected correlation function, , when an ordinary scale-invariant critical point exists.
In the Landau approximation, minimize the Landau free energy to obtain the equation of state, differentiate it with respect to the field conjugate to an order parameter to obtain magnetic susceptibility, and differentiate the minimized free energy with respect to temperature to obtain heat capacity. These operations determine , respectively. Expanding the Landau-Ginzburg theory about its uniform minimum gives the quadratic fluctuation kernel , whose Fourier transform determines the correlation length and the anomalous dimension. At the ordinary quartic critical point this gives and .
For , gives and hence ; the equilibrium free energy gives a finite heat capacity discontinuity, conventionally . Together with the explicit calculation below, the ordinary mean-field critical exponents are
These describe neglect of fluctuation corrections, rather than the exact exponents of every low-dimensional Landau-Ginzburg theory.
A first-order phase transition has a discontinuity in a first derivative of the equilibrium free energy, such as entropy or the order parameter; a thermal first-order transition normally has latent heat. At a continuous phase transition the order parameter goes continuously to zero and there is no latent heat, although response functions and higher derivatives of the free energy can be singular.
For the symmetric scalar Landau free energy at zero conjugate field, minimize
If , the sixth-order term is unnecessary near . For the minimum is , whereas for there are two symmetry-related minima with to leading order. The order parameter vanishes continuously as . The equilibrium free-energy density changes from zero to , so its first temperature derivative is continuous and its second derivative has a jump. This is the ordinary mean-field continuous phase transition.
If , keep for stability. A nonzero stationary order parameter, with , obeys . For phase coexistence with its free energy must also vanish. Substituting into gives
The nonzero solution of is therefore
The competing minima exchange stability at positive , with a finite jump of the order parameter. This is a first-order phase transition. The metastability region is also visible: nonzero stationary points first appear at the spinodal , and the disordered local minimum loses stability at .
For example, holding fixed and taking , the ordered phase has entropy lower than the disordered phase by at coexistence. Heating through the transition absorbs latent heat per volume . More generally the temperature dependence of all coefficients contributes to this entropy difference. With no inversion symmetry, a cubic term can also produce a first-order phase transition, as comparison of the competing minima of the Landau free energy shows.
An order parameter distinguishes thermodynamic phases. For a ferromagnet it is the magnetization; for a fluid near its liquid-gas critical point it can be the density relative to its critical value. In a system with a scalar order parameter, introduce a slowly varying real scalar field whose spatial average represents that order parameter. The Landau-Ginzburg theory organizes the effective free energy as a local expansion consistent with the symmetry:
Here is the stiffness in the gradient energy, and is the field conjugate to an order parameter. The even powers assume a symmetry at . Without that symmetry further terms, including a cubic term, may be allowed. The coefficients are smooth functions of temperature and other controls; near a mean-field transition one usually writes with . A positive highest retained even coefficient makes the local potential bounded below.
The partition function is obtained by weighting configurations with . The Landau approximation replaces this functional average by minimization, taking a uniform in a homogeneous phase. In a symmetric finite system the zero-field average can vanish even below the transition because it averages both signs of . The spontaneous order parameter selects a pure thermodynamic phase by taking the thermodynamic limit before removing the conjugate field:
Thus the order parameter measures the ordering of a selected thermodynamic phase, rather than the average over all symmetry-related phases.
Take the inner product linear in its first argument; for a real space no conjugation convention is needed. Insert into the orthogonality condition against each basis vector . The normal equations are
Equivalently , where and . This orientation of the Gram matrix matches the stated inner-product convention. It is Hermitian and positive definite: for a nonzero coefficient vector ,
by linear independence of the basis. Hence is invertible and gives existence and uniqueness of . For an orthonormal basis this simplifies to ; for a general basis the Gram inverse correctly accounts for interactions between basis vectors.
Write . If is a best approximation in the finite-dimensional linear subspace, then for every in it and every real ,
A nonzero real linear coefficient would make this negative for sufficiently small of the appropriate sign. Hence . In a complex inner product space, apply the argument to too; the imaginary part also vanishes. Thus for every in the subspace.
Conversely, if that orthogonality holds, every other candidate has the form , and the Pythagorean theorem in the inner product norm gives
Equality requires . Therefore
The next part constructs this orthogonal projection explicitly. No completeness of the whole inner product space is required for projection onto a finite-dimensional subspace.
For order six, the support-midpoint sites are , so
By part (b), the complete finite matrix is the symmetric five-diagonal matrix
Only indices from through are retained; the first and last rows are truncated, with no exterior coefficients or wraparound terms. For a row with all possible neighbors, the strict diagonal dominance margin is . Boundary rows have an equal or larger margin.
Here is the appropriate inverse estimate with its proof. If a square matrix has , choose an index maximizing . The reverse triangle inequality gives
Thus implies , and a square finite matrix is invertible. Applying the same estimate with proves . The inverse infinity-norm bound from diagonal dominance consequently yields
This also proves existence and uniqueness of the requested interpolating spline for every data vector and every finite . The quintic cardinal spline midpoint collocation matrix includes the cases of one or two coefficients, where the dominance margin is even larger.
Uniform knots give translation invariance . Put and set missing values to zero. The recurrence in part (a) gives the exact numerical recursion
The order-two starting row is . For example, the order-four middle value is , and the order-six central value is . Iteration gives the requested array, listing :
The symmetry follows from reflection of the uniform Cardinal B-spline about the midpoint of its support. The row sums equal one, consistent with sampling the uniform partition of unity. These are partition-normalized values; normalizing each spline to have peak one would produce a different array.
Use the standard partition normalization of B-splines, with order meaning degree . The Cox-de Boor recurrence on arbitrary increasing spline knots is
starting with . Terms with zero denominator are interpreted as zero in the usual repeated-knot convention. For the uniform knots here, , this reduces to
The index of the second lower-order B-spline is , not . Both lower-order supports combine to give .
The reproduction hypothesis gives , and . If “quadratic” were read as degree exactly two, the same identities follow by subtracting pairs of such reproduced polynomials. Hence
Apply the same positive-barrier argument as in part (a), now with no moment errors. It gives for every and every . Therefore
The identity is itself a positive linear operator on continuous functions reproducing all quadratics, so existence as well as uniqueness is established. In particular, reproducing three small test functions is much stronger for a positive operator than merely fixing one finite-dimensional polynomial space without positivity.
Interpret convergence in as uniform convergence, and work first with real continuous functions. Set , for , and put . The given barriers have one independent of both and . A positive linear operator on continuous functions preserves non-strict inequalities, so application of and evaluation at give
Strict pointwise inequalities need not remain strict under a positive operator; non-strict ones suffice. Linearity and the cancellation of the exact moments yield
Subtracting rather than therefore gives the quadratic barrier estimate for positive approximation operators:
For fixed , all three moment errors tend uniformly to zero. Thus . Since was arbitrary,
This completes the Korovkin theorem proof. Complex-valued functions follow by applying the real result to real and imaginary parts. Uniform convergence of the three test functions is essential to this uniform conclusion; purely pointwise test convergence should not be silently substituted.
Insert a Fourier mode . With and , the amplification polynomial is
Its two roots have product , of modulus one. Thus if both roots are bounded by one, both must have modulus exactly one. Set . The equation becomes
When , both roots have modulus one and their separation obeys for every frequency. The companion matrix has bounded entries and can be diagonalized with the two eigenvectors . The uniform separation bounds the inverse eigenvector matrix. Hence every matrix power is bounded uniformly in and , and the Parseval identity proves L2 norm stability for the two starting levels on the whole grid.
If , frequencies near give one root outside the unit circle, so the scheme is unstable. If , the Nyquist polynomial is , a repeated unit root. The corresponding companion matrix has a nontrivial Jordan block and solutions containing . Thus the endpoints are unstable as well. On the whole line the individual Nyquist plane wave is not square-summable, but Fourier packets localized arbitrarily close to that frequency realize the unbounded power norm by continuity. Therefore the uniform stability of a shifted two-level advection scheme criterion is
In particular is stably exact with compatible starts, whereas is formally exact but unstable under arbitrary perturbations. Checking only the moduli of the roots and retaining both endpoints would miss this essential distinction.
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.
Choose the centered Dirichlet discrete Laplacian in each direction. With put , and . Thus and are the physical derivative approximations; the factor has been absorbed into the Courant number in the written update. Equivalently use the scaled derivative matrices and the time parameter .
For a one-dimensional vector with endpoint values ,
Thus and are real symmetric negative-definite matrices. An orthonormal eigenvector basis shows that all eigenvalues of each exponential lie in for , so
Iterating gives for every nonnegative step parameter, a mesh-independent unconditional stability bound. Multiplying the squared norm by the grid-area weight leaves the same estimate. In this natural tensor-grid choice , so : the split update is even the exact semidiscrete diffusion flow. Contractivity of a product of these negative symmetric exponentials would prove stability even without commutation.
Expand the three matrix exponentials and retain terms through degree two. Their product is
which agrees with to that order. Keeping the cubic terms gives the more informative defect
where is the commutator. For example, the coefficients of in the product are , instead of the exact . The cubic defect is generically nonzero. Therefore Strang splitting has order . The symmetry is consistent with this cancellation of the quadratic defect; for commuting the composition is exact rather than merely second order.

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