Differentiate the Airy ordinary differential equation to obtain . Combining part (a)'s derivatives yields
To make this vanish for an arbitrary Airy profile, the time powers must agree: , giving . Their common coefficient must then vanish, giving , hence . Therefore
For the two terms cancel exactly:
This is an Airy similarity solution of the linear dispersive equation. The Airy equation here is the linear partial differential equation , the reduced equation specified in the question; it is distinct from the profile's Airy ordinary differential equation. The construction works for every linear combination of its two independent profiles, including zero. No condition at is supplied, and this singular time is excluded. With real cube roots the final formula can also be used separately on .
For this separable differential equation, away from , separation gives , or on a connected interval. The initial condition gives , hence
The maximal interval containing on which this solves the printed differential equation is . Although the linear formula extends through one, the right-hand side there is and the original equation is undefined. The equilibrium solution , lost if one divides by , does not satisfy the initial condition.
Differentiate the Fourier series term by term and collect cosine and sine terms. Substitution into the equation gives
For a smooth periodic solution, differentiation of the series is justified. Alternatively, multiply the differential equation by each trigonometric mode and integrate by parts; periodicity cancels the boundary terms and gives the same coefficient constraints without first differentiating an infinite sum.
Construct the scalar potential . Its differential is
which proves exactness directly. Multiplying the differential equation by shows that a solution satisfies , so . The initial value at zero gives . Therefore
The factor is also an integrating factor for the associated first-order linear equation.
First solve for the velocity components. The integrating factor gives
Integrating once more and imposing the initial position gives
These formulas describe projectile motion with linear drag. Differentiation checks both second-order differential equations and both initial velocities. At large positive time, the horizontal velocity tends to zero and the vertical velocity tends to the terminal value .
Let and . Their cross partial derivatives agree: . More directly, construct the scalar function
This exhibits the expression as an exact differential. Along a solution of the differential equation,
Thus the general solution is
For , the graph also extends smoothly through and satisfies the original, undivided equation there. Nonzero cannot extend through zero as a finite-valued differentiable solution. The extra vertical component of the implicit zero-level equation is not a graph on an interval and is not an additional solution of this ordinary differential equation.
With , integrate the coefficient with respect to :
Its other partial derivative is . Matching the coefficient gives . Therefore the differential equation is , and its implicit general solution is
These are level curves of the potential of a conservative vector field. Where , they give the usual graph solution . At , the implicit description retains possible vertical tangents instead of dividing by zero.
The nonzero factor can be cancelled in the local differential equation, leaving . On an interval with ,
Hence the explicit general solution is
on any interval that avoids the excluded origin. It also follows from fixing the local angle potential, with slope . As parameterized curves there are also vertical rays , which cannot be represented as single-valued functions . Geometrically the connected integral curves are rays, not lines continued through the undefined origin.
The original differential equation is defined only where and . Put , so the chain rule gives . Multiplying the equation by yields a first-order linear ordinary differential equation:
An integrating factor on either half-line is , up to an irrelevant nonzero constant. Equivalently, apply the quotient rule directly:
Integrating gives . Thus the real general solution is
on any connected interval where and . These restrictions preserve both the original denominator and differentiability of the real cube root. In particular includes on either half-line; this solution has not been lost through division by an expression involving .
For verification, differentiating gives , while . The two sides agree wherever the original equation is defined.
Use , so . Up to an overall minus sign, the differential form becomes
Seeking a monomial integrating factor , equality of mixed derivatives gives
These give , so the integrating factor in the original variables is
One may verify exactness directly without relying on invertibility of the cosine substitution:
Thus the implicit first integral is
The equivalence with the differential equation holds directly where . The zero solution is included at . At a zero of , the original graph equation requires for any finite differentiable continuation, so no extra finite graph with nonzero is created by multiplying by the vanishing factor. If one allows vertical parametrized integral curves of the differential form, is also such a curve and is already contained in the level set .
Write the differential form as , with the coefficients as printed. A monomial integrating factor can be found from . The exactness condition gives
Solving yields . Thus, on ,
After multiplication, the differential form is
Hence the implicit solutions are
The branch was excluded by the integrating factor but satisfies the original equation and must be restored. On any graph branch with , differentiation of the boxed first integral recovers the original differential equation.
On an interval excluding , an integrating factor for this first-order linear differential equation is . Multiplying through gives
An antiderivative then gives , so
The constant is arbitrary on each connected interval of the domain; the original differential equation is undefined at zero.
Because the original differential equation divides by , its domain requires and . Set , so . The equation becomes the first-order linear differential equation
An integrating factor is , since . Thus
Returning to gives the explicit general real solutions
Each sign is fixed on a connected interval where and the radicand is strictly positive. The strict condition excludes , where the original differential equation is undefined. There is no additional zero solution.
Where , divide by it to obtain . The integrating factor is , and the product rule gives
The initial condition gives . Hence
The simplified expression has a differentiable extension through , with . Direct substitution verifies the original undivided differential equation there as well: its left side is . Thus the maximal interval containing on which this solution is finite is ; the apparent exclusion of arose only from division by . This is continuation through a zero of the leading ODE coefficient.
On an interval with , division by gives the separable differential equation . For , take antiderivatives:
Renaming the constant yields the explicit nonzero family
For , the interval of definition must avoid zeros of the denominator. Division by excluded the additional solution , which must also be included.
The original undivided differential equation also makes sense at , where it requires . For continuation through a zero of the leading ODE coefficient, rewrite the nonzero family as
It extends differentiably with and , since . One may choose the integration constants independently on the two sides of zero and obtain a differentiable solution through zero. The zero solution also extends through zero, with derivative ; joining a nonzero branch on just one side to a zero branch on the other would fail differentiability.
Again , but the principal inverse sine would give at . The required branch is , corresponding to . Therefore
At the bracket is , as required. Both the initial condition and the original differential equation are thus satisfied on the stated branch.
For a nonzero real function, write with . Multiplying the Sturm-Liouville equation by and integrating gives
Thus the stated vanishing boundary product proves .
For the variational statement, the first variation is
With admissible boundary conditions that also make for every allowed variation, the fundamental lemma of the calculus of variations identifies stationary functions with the eigenfunctions and stationary values with their eigenvalues. Zero Dirichlet boundary conditions, zero Neumann boundary conditions, and matched periodic data give familiar examples. The single printed condition proves the quotient identity but, by itself, does not specify the admissible variations needed for stationarity; the boundary conditions for variational Rayleigh quotients make this distinction explicit.
For the half-line equation, multiplication by gives and . Use the regular radial boundary condition at zero and square-integrable decay at infinity, under which the proposed Gaussian is admissible. With , , , so
The ratio of the two Gaussian integrals is , from a change of variables and the Gamma function recurrence. Differentiating gives , so the minimum on occurs at . The Gaussian variational bound for a radial harmonic oscillator therefore gives
Here the trial actually satisfies the differential equation with . Moreover, on the same regular decaying domain, integration by parts gives . Hence the estimate is exact, with . For , regular radial data mean the Neumann boundary condition ; imposing a zero Dirichlet boundary condition there would exclude this trial and give a different problem.
The origin is an ordinary point of the Hermite differential equation, so put . Substitution and matching powers gives
Thus the linear recurrence relation is
Even and odd coefficients evolve independently from arbitrary . The general power series solution is
with an empty product equal to one. Its first terms are
Both series converge at every finite , since the ratio of successive terms in each parity subsequence tends to zero. For a nonzero polynomial of degree , the highest coefficient in the differential equation requires . Conversely, at such a value the appropriate parity series terminates, while its opposite-parity initial coefficient can be set to zero. Hence
Here is a physicists' Hermite polynomial. Two explicit choices are
Direct differentiation verifies both. These are polynomial solutions up to an arbitrary overall factor. Without a polynomial requirement or additional boundary conditions, the equation also has nonpolynomial solutions for other real ; discreteness here comes from termination.
Multiplication by a positive weight puts the differential equation into Sturm-Liouville form if . Thus , and the constant factor can be chosen so that
This fixes the otherwise arbitrary normalization of the weight consistently with the requested unit norms.
Write the polynomial as . Comparing coefficients gives
Consequently , identifying the solutions as Laguerre polynomials. In particular,
Repeated integration by parts gives the moments . Therefore
For later use, integration by parts in the weighted eigenvalue equations also gives
The endpoint expression vanishes because of the factor at zero and exponential decay at infinity. Hence distinct Laguerre polynomials are orthogonal, an instance of orthogonality of Sturm-Liouville eigenfunctions.
Differentiate the power series term by term inside its radius of convergence, and reindex the second derivative so that every coefficient multiplies . The differential equation becomes
The uniqueness of power-series coefficients gives
The even coefficients depend only on and the odd coefficients only on . These are the two local solution branches of the Legendre differential equation around its ordinary point .