The Central Idea
An ordinary differential equation may have many solutions. Boundary conditions restrict that family by requiring the solution to satisfy specified behavior at selected locations.
For example, consider
\[ y''+\lambda y=0, \]together with
\[ y(0)=0, \qquad y(L)=0. \]The differential equation permits many possible functions, but the endpoint conditions allow nonzero solutions only for special values of \(\lambda\). Those values become eigenvalues, and the associated solutions become eigenfunctions.
Boundary conditions therefore do more than choose constants. In many problems, they determine the spectral structure of the system itself.
What Is a Boundary-Value Problem?
A boundary-value problem, or BVP, consists of:
- a differential equation defined on an interval or region;
- conditions imposed at the boundary of that interval or region;
- a requirement to find a function satisfying both the equation and the boundary conditions.
A typical second-order BVP on \([a,b]\) has the form
\[ y''+p(x)y'+q(x)y=r(x), \]together with two independent boundary conditions involving values or derivatives of \(y\) at \(a\) and \(b\).
Boundary Conditions versus Initial Conditions
An initial-value problem typically specifies all required data at one point. For example,
\[ y''+y=0, \qquad y(0)=1, \qquad y'(0)=0. \]A boundary-value problem often distributes conditions across different points:
\[ y''+y=0, \qquad y(0)=0, \qquad y(\pi)=0. \]This distinction changes the mathematics. Initial-value problems are often solved by evolving information forward from one location. Boundary-value problems must satisfy constraints at multiple locations simultaneously.
Common Types of Boundary Conditions
Several standard boundary conditions recur across applications.
Dirichlet Conditions
The value of the function is specified at the boundary:
\[ y(a)=A, \qquad y(b)=B. \]Neumann Conditions
The derivative is specified:
\[ y'(a)=C, \qquad y'(b)=D. \]Robin Conditions
A linear combination of the function and its derivative is specified:
\[ \alpha y(a)+\beta y'(a)=C. \]Robin conditions model many mixed physical effects, including convective heat transfer and elastic restraints.
Homogeneous and Nonhomogeneous Boundary Conditions
Boundary conditions are homogeneous when the prescribed boundary values are zero. For example,
\[ y(0)=0, \qquad y(L)=0. \]Conditions such as
\[ y(0)=3, \qquad y(L)=5 \]are nonhomogeneous.
Homogeneous boundary conditions are especially important in spectral methods because they often produce clean eigenvalue problems. Nonhomogeneous conditions can frequently be converted into homogeneous ones by subtracting a simple function that already satisfies the prescribed boundary values.
Example: A Simple Dirichlet Problem
Solve
\[ y''=0, \qquad y(0)=2, \qquad y(4)=10. \]Integrating twice gives
\[ y(x)=C_1x+C_2. \]From \(y(0)=2\),
\[ C_2=2. \]From \(y(4)=10\),
\[ 4C_1+2=10, \]so
\[ C_1=2. \]Therefore,
\[ \boxed{y(x)=2x+2}. \]In this elementary example, the two boundary conditions determine the two arbitrary constants in the general solution.
Existence and Uniqueness Are Not Automatic
Boundary-value problems behave differently from many familiar initial-value problems. Depending on the equation and boundary conditions, a BVP may have:
- exactly one solution;
- no solution;
- infinitely many solutions.
The boundary conditions can be incompatible, redundant, or aligned with an eigenvalue of the associated homogeneous operator. Recognizing these possibilities is an important part of BVP analysis.
Example: No Solution
Consider
\[ y''=0, \qquad y'(0)=1, \qquad y'(2)=3. \]Since \(y''=0\), the derivative must be constant:
\[ y'(x)=C_1. \]It cannot equal both \(1\) and \(3\). Therefore, this boundary-value problem has no solution.
Example: Infinitely Many Solutions
Consider
\[ y''+y=0, \qquad y(0)=0, \qquad y(\pi)=0. \]The general solution is
\[ y(x)=A\cos x+B\sin x. \]The first condition gives \(A=0\). The second gives
\[ B\sin\pi=0, \]which is true for every value of \(B\). Thus,
\[ \boxed{y(x)=B\sin x} \]is a family of solutions. The loss of uniqueness occurs because \(1\) is an eigenvalue of the associated boundary-value operator.
Boundary Conditions Create Eigenvalue Problems
Consider the parameterized problem
\[ y''+\lambda y=0, \qquad y(0)=0, \qquad y(L)=0. \]We seek values of \(\lambda\) for which a nontrivial solution exists.
When \(\lambda>0\), write
\[ \lambda=\mu^2. \]Then
\[ y(x)=A\cos(\mu x)+B\sin(\mu x). \]The condition \(y(0)=0\) gives \(A=0\). The condition \(y(L)=0\) requires
\[ B\sin(\mu L)=0. \]For a nonzero solution, \(B\neq0\), so
\[ \sin(\mu L)=0. \]Hence,
\[ \mu L=n\pi, \qquad n=1,2,3,\ldots \]and therefore
\[ \boxed{\lambda_n=\left(\frac{n\pi}{L}\right)^2}. \]The Corresponding Eigenfunctions
For each eigenvalue
\[ \lambda_n=\left(\frac{n\pi}{L}\right)^2, \]the corresponding eigenfunction is
\[ \boxed{\phi_n(x)=\sin\left(\frac{n\pi x}{L}\right)}. \]These functions satisfy the endpoint conditions automatically and are mutually orthogonal on \([0,L]\):
\[ \int_0^L \sin\left(\frac{m\pi x}{L}\right) \sin\left(\frac{n\pi x}{L}\right) \,dx=0, \qquad m\neq n. \]This is the structural reason sine series appear naturally in many boundary-value problems.
The Sturm-Liouville Form
A broad class of linear eigenvalue problems can be written in Sturm-Liouville form:
\[ -\frac{d}{dx} \left(p(x)\frac{dy}{dx}\right) +q(x)y =\lambda w(x)y. \]Here:
- \(p(x)\) controls the derivative term;
- \(q(x)\) acts like a potential or restoring term;
- \(w(x)\) is a positive weight function;
- \(\lambda\) is the eigenvalue parameter.
Appropriate boundary conditions make the differential operator self-adjoint, which leads to real eigenvalues and orthogonal eigenfunctions.
Weighted Orthogonality
In a Sturm-Liouville problem, distinct eigenfunctions satisfy a weighted orthogonality condition:
\[ \boxed{ \int_a^b w(x)\phi_m(x)\phi_n(x)\,dx=0, \qquad m\neq n }. \]The weight \(w(x)\) belongs to the geometry of the problem. It determines the inner product with respect to which the eigenfunctions are orthogonal.
From Boundary Conditions to Expansion Bases
Once the admissible eigenfunctions are known, a suitable function \(f(x)\) can often be expanded as
\[ f(x)=\sum_{n=1}^{\infty}c_n\phi_n(x). \]If the eigenfunctions are orthogonal under weight \(w(x)\), then
\[ \boxed{ c_n= \frac{ \int_a^b f(x)\phi_n(x)w(x)\,dx }{ \int_a^b \phi_n(x)^2w(x)\,dx } }. \]The boundary-value problem therefore creates the basis in which the associated differential operator becomes spectrally simple.
Why Boundary Conditions Matter Spectrally
The same differential expression can produce different spectra under different boundary conditions. For example, the operator
\[ -\frac{d^2}{dx^2} \]on \([0,L]\) has different admissible modes depending on whether we impose Dirichlet, Neumann, periodic, or mixed conditions.
Boundary conditions are therefore part of the operator specification. Changing them can change the eigenvalues, eigenfunctions, and physical interpretation of the model.
Dirichlet versus Neumann Spectra
For Dirichlet conditions
\[ y(0)=y(L)=0, \]the standard eigenfunctions are
\[ \sin\left(\frac{n\pi x}{L}\right), \qquad n=1,2,\ldots \]For Neumann conditions
\[ y'(0)=y'(L)=0, \]cosine modes appear:
\[ \cos\left(\frac{n\pi x}{L}\right), \qquad n=0,1,2,\ldots \]The \(n=0\) Neumann mode is constant, reflecting the fact that a constant function has zero derivative at both endpoints.
Mixed Boundary Conditions
Suppose
\[ y(0)=0, \qquad y'(L)=0. \]For
\[ y''+\lambda y=0, \]the first condition selects sine-type solutions. The second requires
\[ \cos(\mu L)=0. \]Thus,
\[ \mu L=\left(n+\frac12\right)\pi, \qquad n=0,1,2,\ldots \]and
\[ \boxed{ \lambda_n= \left(\frac{(n+\frac12)\pi}{L}\right)^2 }. \]Mixed conditions shift the spectrum away from the ordinary integer harmonics.
Nonhomogeneous Boundary Conditions by Substitution
Suppose a function satisfies
\[ u(0,t)=A, \qquad u(L,t)=B. \]Choose a simple function \(v(x)\) satisfying the boundary values, such as
\[ v(x)=A+\frac{B-A}{L}x. \]Write
\[ u(x,t)=v(x)+w(x,t). \]Then
\[ w(0,t)=w(L,t)=0. \]This converts the boundary conditions into homogeneous form. The differential equation for \(w\) may acquire an additional forcing term, but the spectral basis becomes much easier to use.
Separation of Variables
Boundary-value problems become especially important for partial differential equations. Consider the heat equation
\[ u_t=\kappa u_{xx}, \qquad 0Seek a separated solution
\[ u(x,t)=X(x)T(t). \]Substitution gives
\[ X(x)T'(t)=\kappa X''(x)T(t). \]Dividing by \(\kappa XT\) yields
\[ \frac{T'}{\kappa T} = \frac{X''}{X} =-\lambda. \]The PDE has separated into a time equation and a spatial boundary-value problem.
The Spatial Eigenvalue Problem
Separation gives
\[ X''+\lambda X=0, \]with
\[ X(0)=X(L)=0. \]Thus,
\[ \lambda_n=\left(\frac{n\pi}{L}\right)^2, \]and
\[ X_n(x)=\sin\left(\frac{n\pi x}{L}\right). \]The time equation becomes
\[ T_n'(t)+\kappa\lambda_nT_n(t)=0, \]so
\[ T_n(t)=e^{-\kappa\lambda_n t}. \]Heat Equation Solution by Eigenfunction Expansion
Combining the separated modes gives
\[ \boxed{ u(x,t)= \sum_{n=1}^{\infty} b_n e^{-\kappa(n\pi/L)^2t} \sin\left(\frac{n\pi x}{L}\right) }. \]If the initial temperature is
\[ u(x,0)=f(x), \]then the coefficients are the sine-series coefficients of \(f\):
\[ \boxed{ b_n= \frac{2}{L} \int_0^L f(x) \sin\left(\frac{n\pi x}{L}\right) \,dx }. \]Each spatial eigenmode evolves independently in time.
Why Higher Heat Modes Decay Faster
The \(n\)-th heat mode contains the factor
\[ e^{-\kappa(n\pi/L)^2t}. \]Since the exponent contains \(n^2\), higher-frequency spatial modes decay more quickly. Fine-scale temperature variation disappears before broad-scale variation.
This is a spectral explanation of diffusion: the heat equation behaves like a low-pass filter in space.
The Wave Equation
Consider
\[ u_{tt}=c^2u_{xx}, \qquad 0The same spatial boundary-value problem produces the same eigenfunctions:
\[ \sin\left(\frac{n\pi x}{L}\right). \]But the time equation is oscillatory rather than decaying. The general modal form is
\[ u(x,t)= \sum_{n=1}^{\infty} \left[ A_n\cos\left(\frac{cn\pi t}{L}\right) + B_n\sin\left(\frac{cn\pi t}{L}\right) \right] \sin\left(\frac{n\pi x}{L}\right). \]The boundary conditions select the standing-wave frequencies.
Steady-State Boundary-Value Problems
Some PDE problems become boundary-value problems after time dependence disappears. For steady one-dimensional heat conduction without internal sources,
\[ u_{xx}=0. \]With
\[ u(0)=T_0, \qquad u(L)=T_L, \]the steady temperature is
\[ \boxed{ u(x)=T_0+ rac{T_L-T_0}{L}x }. \]Boundary values alone determine the linear equilibrium profile.
Poisson-Type Boundary-Value Problems
A source term produces equations such as
\[ -u''(x)=f(x), \qquad u(0)=u(L)=0. \]The forcing \(f(x)\) drives the system while the boundary conditions constrain the response. Such problems can be solved by direct integration in simple cases, by eigenfunction expansions, by Green's functions, or by numerical discretization.
Green's Functions
A Green's function describes the response of a boundary-value operator to a point source. For an operator \(L\), the Green's function \(G(x,\xi)\) is chosen so that
\[ L_xG(x,\xi)=\delta(x-\xi), \]together with the required boundary conditions in the variable \(x\).
Then a forced problem
\[ Lu=f \]can often be written
\[ \boxed{ u(x)=\int_a^bG(x,\xi)f(\xi)\,d\xi }. \]The Green's function acts as the impulse response of the boundary-value operator.
Green's Functions and Spectral Decomposition
Suppose a self-adjoint operator has normalized eigenfunctions \(\phi_n\) with nonzero eigenvalues \(\lambda_n\). Formally, the inverse operator can be expanded as
\[ G(x,\xi) = \sum_n \frac{\phi_n(x)\phi_n(\xi)}{\lambda_n}. \]This formula reveals a direct connection among boundary conditions, eigenfunction expansions, spectral decomposition, and Green's functions. Each spectral mode is weighted by the reciprocal of its eigenvalue.
Resonance and Compatibility
Problems become delicate when the forcing aligns with a zero mode or with the homogeneous solution space. In such cases, a solution may fail to exist unless a compatibility condition is satisfied.
For example, consider the Neumann problem
\[ -u''(x)=f(x), \qquad u'(0)=u'(L)=0. \]Integrating over \([0,L]\) gives
\[ -\int_0^L u''(x)\,dx = -u'(L)+u'(0) =0. \]Therefore, a necessary compatibility condition is
\[ \boxed{ \int_0^L f(x)\,dx=0 }. \]The constant Neumann eigenfunction is a zero mode, so the forcing must have zero projection onto that mode.
Transform Methods for Boundary Problems
Fourier sine and cosine transforms are natural for problems on half-lines with suitable endpoint conditions. Full Fourier transforms are useful on unbounded domains. Laplace transforms are especially useful when time is present and initial conditions must be incorporated algebraically.
Transforming one variable often converts a PDE into an ordinary differential boundary-value problem in the remaining variable.
Example: Laplace Transform in Time
For the heat equation
\[ u_t=\kappa u_{xx}, \]define
\[ U(x,s)=\mathcal{L}\{u(x,t)\}. \]If \(u(x,0)=f(x)\), then
\[ \mathcal{L}\{u_t\}=sU-f(x). \]The transformed equation is
\[ \kappa U_{xx}-sU=-f(x). \]For each transform parameter \(s\), we now have an ordinary differential boundary-value problem in \(x\).
Self-Adjoint Operators and Boundary Terms
Orthogonality results often follow from integration by parts. For a simple second-derivative operator,
\[ \int_a^b u(-v'')\,dx - \int_a^b (-u'')v\,dx \]reduces to boundary terms:
\[ \left[-uv'+u'v\right]_a^b. \]Suitable boundary conditions make this expression vanish. That is why boundary conditions are central to self-adjointness and, consequently, to real spectra and orthogonal eigenfunctions.
Boundary Conditions as Physical Constraints
Boundary conditions encode physical information. Examples include:
- a string fixed at both ends;
- an insulated boundary with zero heat flux;
- a prescribed temperature at a surface;
- a beam clamped or free at an endpoint;
- a fluid velocity constrained at a wall;
- an electric potential prescribed on a conductor;
- a convective boundary coupling temperature and heat flux.
Choosing the wrong mathematical boundary condition can therefore change the model, not merely the method of solution.
Higher-Dimensional Boundary-Value Problems
In two or three dimensions, the boundary is a curve or surface rather than a pair of endpoints. A classical example is Laplace's equation
\[ \nabla^2u=0 \]on a region \(\Omega\), together with specified values of \(u\) on the boundary \(\partial\Omega\).
Geometry matters strongly. Rectangular domains naturally lead to sine and cosine products, circular domains to trigonometric and Bessel functions, and spherical domains to spherical harmonics.
Boundary Geometry Selects the Spectral Basis
Spectral functions are not arbitrary decorations. They arise from the geometry and boundary constraints of the problem.
- intervals lead to trigonometric modes;
- cylindrical symmetry leads to Bessel functions;
- spherical symmetry leads to spherical harmonics and related radial functions;
- general domains lead to eigenfunctions of the corresponding differential operator.
The correct basis is therefore adapted to the operator, the domain, and the boundary conditions simultaneously.
Numerical Boundary-Value Problems
Many BVPs cannot be solved in elementary closed form. Common numerical strategies include:
- shooting methods;
- finite-difference methods;
- finite-element methods;
- collocation methods;
- spectral methods.
Each method converts the differential problem into a finite computational problem, but the boundary conditions must still be enforced correctly.
The Shooting Method
A shooting method converts a boundary-value problem into a sequence of initial-value problems. Suppose
\[ y''=F(x,y,y'), \qquad y(a)=A, \qquad y(b)=B. \]The unknown initial slope
\[ y'(a)=s \]is guessed. The initial-value problem is integrated to \(x=b\), and the guess is adjusted until
\[ y(b;s)=B. \]The method reduces the BVP to a root-finding problem in the shooting parameter \(s\).
Finite-Difference Discretization
For
\[ -u''(x)=f(x), \qquad u(0)=u(L)=0, \]choose grid points \(x_i=ih\). At an interior point,
\[ u''(x_i) \approx \frac{u_{i-1}-2u_i+u_{i+1}}{h^2}. \]Therefore,
\[ \frac{-u_{i-1}+2u_i-u_{i+1}}{h^2} \approx f_i. \]Applying this relation at every interior grid point produces a linear system whose matrix is a discrete version of the differential operator.
Discrete Spectra of Finite-Difference Operators
The finite-difference matrix for a second derivative with Dirichlet conditions has eigenvectors that resemble sampled sine waves. As the grid is refined, its eigenvalues and eigenvectors approximate the continuous boundary-value spectrum.
This provides a direct bridge between continuous eigenfunction expansions and numerical linear algebra.
Spectral Methods
A spectral method approximates the unknown solution directly in a global basis:
\[ u_N(x)=\sum_{n=0}^{N}c_n\phi_n(x). \]The basis functions are often chosen to reflect the geometry and boundary conditions. Fourier modes, Chebyshev polynomials, Legendre polynomials, and operator eigenfunctions are common choices.
For sufficiently smooth solutions, spectral methods can converge extremely rapidly.
A Reliable Workflow for Boundary-Value Problems
- Identify the differential equation and its domain.
- Classify the boundary conditions as Dirichlet, Neumann, Robin, periodic, or mixed.
- Determine whether the boundary conditions are homogeneous.
- If needed, transform nonhomogeneous boundary conditions into homogeneous ones.
- Solve the associated homogeneous differential equation.
- Apply the boundary conditions carefully.
- Check whether special parameter values create eigenfunctions.
- Use orthogonality or transforms when a spectral expansion is appropriate.
- Check compatibility conditions, especially when zero modes are present.
- Verify both the differential equation and every boundary condition.
Checking a Boundary-Value Solution
A candidate solution must pass two independent tests.
- Substitute the function into the differential equation.
- Evaluate the function or derivatives at every boundary and verify each condition.
Satisfying only the differential equation is not enough. Satisfying only the boundary conditions is not enough. A BVP solution must satisfy the complete system.
Common Error: Treating a BVP Like an IVP
Boundary conditions may be located at different points. One cannot generally apply an initial-value formula without accounting for the condition at the far boundary.
The correct approach is to retain enough unknown constants or parameters to enforce all boundary constraints simultaneously.
Common Error: Ignoring the Zero Solution
In an eigenvalue problem, the zero function satisfies homogeneous boundary conditions for every \(\lambda\). But an eigenfunction must be nonzero.
The goal is therefore to determine which parameter values permit a nontrivial solution.
Common Error: Dividing by a Quantity That May Vanish
During eigenvalue analysis, an expression such as
\[ B\sin(\mu L)=0 \]must not be simplified by dividing by \(B\) unless the nontrivial-solution condition \(B\neq0\) has already been stated. Likewise, one should not divide by \(\sin(\mu L)\), because its zeros are exactly what determine the eigenvalues.
Common Error: Using the Wrong Eigenfunctions
A sine basis is natural for homogeneous Dirichlet conditions, while a cosine basis is natural for homogeneous Neumann conditions. Using a basis that violates the boundary conditions creates unnecessary work and may lead to an invalid expansion.
The boundary conditions should guide the choice of spectral basis from the start.
Common Error: Forgetting Compatibility Conditions
When an operator has a zero eigenvalue, the inverse operator does not exist on the entire function space. The forcing must be orthogonal to the corresponding null space for a solution to exist.
Neumann Poisson problems provide a standard example of this issue.
Common Error: Applying Boundary Conditions to the Wrong Variable
In separation of variables, a physical boundary condition on \(u(x,t)\) usually becomes a condition on the spatial factor \(X(x)\), not directly on the time factor \(T(t)\).
Keeping track of which independent variable belongs to each condition prevents many setup errors.
Practice 1: Elementary Dirichlet Problem
Solve
\[ y''=0, \qquad y(0)=1, \qquad y(3)=7. \]Practice 2: Identify the Boundary Type
Classify each condition.
- \(y(0)=4\)
- \(y'(L)=0\)
- \(2y(L)+3y'(L)=5\)
Practice 3: Dirichlet Eigenvalues
For
\[ y''+\lambda y=0, \qquad y(0)=0, \qquad y(2)=0, \]find the positive eigenvalues and corresponding eigenfunctions.
Practice 4: Neumann Zero Mode
Show that \(\lambda=0\) is an eigenvalue of
\[ -y''=\lambda y, \qquad y'(0)=y'(L)=0. \]Identify an eigenfunction.
Practice 5: Mixed Conditions
For
\[ y''+\lambda y=0, \qquad y(0)=0, \qquad y'(L)=0, \]determine the positive eigenvalues.
Practice 6: Homogenize the Boundary Conditions
Suppose
\[ u(0,t)=2, \qquad u(L,t)=8. \]Find a simple function \(v(x)\) such that \(w=u-v\) satisfies homogeneous Dirichlet conditions.
Practice 7: Heat-Equation Mode
For a rod of length \(L\) with zero endpoint temperatures, determine the time dependence of the spatial mode
\[ \sin\left(\frac{3\pi x}{L}\right) \]under the heat equation \(u_t=\kappa u_{xx}\).
Practice 8: Compatibility Condition
Determine whether the Neumann problem
\[ -u''=1, \qquad u'(0)=u'(L)=0 \]can have a solution for \(L>0\).
Practice 9: Finite-Difference Equation
Write the centered finite-difference approximation to
\[ -u''(x_i)=f(x_i) \]using grid spacing \(h\).
Practice 10: Green's-Function Interpretation
In one sentence, explain what \(G(x,\xi)\) represents in the formula
\[ u(x)=\int_a^bG(x,\xi)f(\xi)\,d\xi. \]Practice Answers
Practice 1
Since \(y''=0\),
\[ y=C_1x+C_2. \]The condition \(y(0)=1\) gives \(C_2=1\). Then
\[ 3C_1+1=7, \]so \(C_1=2\). Therefore,
\[ \boxed{y(x)=2x+1}. \]Practice 2
- \(y(0)=4\): Dirichlet.
- \(y'(L)=0\): Neumann.
- \(2y(L)+3y'(L)=5\): Robin.
Practice 3
Here \(L=2\), so
\[ \lambda_n=\left(\frac{n\pi}{2}\right)^2, \qquad n=1,2,3,\ldots \]with eigenfunctions
\[ \boxed{ \phi_n(x)=\sin\left(\frac{n\pi x}{2}\right) }. \]Practice 4
When \(\lambda=0\), the equation becomes
\[ y''=0, \]so \(y=Ax+B\). The Neumann conditions require \(A=0\), leaving
\[ \boxed{y(x)=B}, \]where \(B\neq0\). Thus a nonzero constant function is an eigenfunction for the zero eigenvalue.
Practice 5
The mixed boundary conditions require
\[ \cos(\mu L)=0. \]Hence
\[ \mu_n=\frac{(n+\frac12)\pi}{L}, \qquad n=0,1,2,\ldots \]and
\[ \boxed{ \lambda_n= \left(\frac{(n+\frac12)\pi}{L}\right)^2 }. \]Practice 6
A linear function connecting the boundary values is
\[ \boxed{ v(x)=2+\frac{6}{L}x }. \]Then \(w=u-v\) satisfies
\[ w(0,t)=w(L,t)=0. \]Practice 7
The third spatial eigenvalue is
\[ \lambda_3=\left(\frac{3\pi}{L}\right)^2. \]Therefore the mode decays as
\[ \boxed{ e^{-\kappa(3\pi/L)^2t} \sin\left(\frac{3\pi x}{L}\right) }. \]Practice 8
The compatibility condition requires
\[ \int_0^L1\,dx=L=0. \]This is impossible for \(L>0\). Therefore the BVP has no solution.
Practice 9
Using the centered second difference,
\[ \boxed{ \frac{-u_{i-1}+2u_i-u_{i+1}}{h^2}=f_i }. \]Practice 10
The Green's function \(G(x,\xi)\) is the response observed at \(x\) to a unit point source applied at \(\xi\), subject to the specified boundary conditions.
What to Remember
A boundary-value problem combines a differential equation with constraints imposed at the boundary of its domain.
- Dirichlet conditions prescribe function values;
- Neumann conditions prescribe normal derivatives or fluxes;
- Robin conditions prescribe linear combinations of values and derivatives;
- boundary conditions may produce one, no, or infinitely many solutions;
- homogeneous boundary conditions naturally lead to eigenvalue problems;
- different boundary conditions can produce different spectra for the same differential expression;
- Sturm-Liouville systems provide real eigenvalues and weighted orthogonal eigenfunctions under appropriate assumptions;
- eigenfunction expansions convert differential operators into mode-by-mode algebra;
- separation of variables turns many PDEs into spatial boundary-value eigenproblems;
- heat modes decay according to their eigenvalues;
- wave modes oscillate at frequencies determined by the same spatial spectrum;
- Green's functions represent inverse boundary-value operators as responses to point forcing;
- zero modes can create compatibility conditions;
- finite-difference matrices approximate continuous boundary-value operators and their spectra;
- spectral methods approximate solutions directly in boundary-adapted global bases.
The central structural lesson is that a differential equation, its domain, and its boundary conditions form a single mathematical system. The spectral behavior cannot be understood correctly by considering the differential expression alone.
Where We Go Next
The transform and spectral tools developed throughout this section are powerful in exact analysis, but real applications usually require finite data, numerical quadrature, sampled transforms, matrix eigenproblems, truncation, interpolation, and efficient algorithms. The next step is to turn the analytical framework into a computational workflow.
The next lesson, Computational Transform Methods, develops numerical implementations of Fourier, Laplace, convolution, and spectral procedures. We will examine discretization, FFT-based computation, transform grids, numerical inversion, truncation and aliasing, matrix-based spectral approximations, conditioning, and practical error control.