Mathematics / Differential Equations Canonical Reductions & Green's Functions 100% Free Open Access
Chapter 4 • Theory & Derivations

Unit 4: The 1D & Multi-D Wave Equation: D'Alembert's Formula & Energy Methods

Exhaustive treatment of hyperbolic wave propagation: physical derivation from tensioned strings and acoustics, D'Alembert's traveling wave formula for infinite strings, domain of dependence, range of influence and relativistic causality cones, semi-infinite strings and the method of images, Fourier separation of variables into standing normal modes, and rigorous proofs of energy conservation and solution uniqueness via energy integrals.

§4.1 The 1D Infinite String Cauchy Problem: D'Alembert's Formula

1. Physical Derivation of the 1D Wave Equation

Consider a flexible, perfectly elastic string of constant linear mass density $\rho$ stretched under high uniform tension $T$ along the $x$-axis. Let $u(x, t)$ denote the small transverse displacement at position $x$ and time $t$. Assuming small slopes ($|\partial u/\partial x| \ll 1$), tension variations are negligible, and longitudinal displacements are negligible.

Consider an infinitesimal string element between $x$ and $x + \Delta x$. The vertical component of tension at $x + \Delta x$ is $T \sin\theta_2 \approx T \tan\theta_2 = T u_x(x + \Delta x, t)$. The vertical component of tension at $x$ is $-T \sin\theta_1 \approx -T u_x(x, t)$. By Newton's second law ($F = m a$):

$$\Delta m \frac{\partial^2 u}{\partial t^2} = (\rho \Delta x) u_{tt} = T [u_x(x + \Delta x, t) - u_x(x, t)]$$

Dividing by $\rho \Delta x$ and taking the limit $\Delta x \to 0$:

$$\frac{\partial^2 u}{\partial t^2} = \frac{T}{\rho} \frac{\partial^2 u}{\partial x^2}$$

Setting the propagation wave speed $c = \sqrt{T/\rho}$ yields the 1D wave equation:

$$u_{tt} - c^2 u_{xx} = 0$$

2. Complete Derivation of D'Alembert's Formula

Consider the Cauchy initial value problem on the infinite domain $x \in \mathbb{R}$:

$$\begin{cases} u_{tt} - c^2 u_{xx} = 0, & x \in \mathbb{R}, \quad t > 0 \\ u(x, 0) = f(x), & x \in \mathbb{R} \quad (\text{initial displacement}) \\ u_t(x, 0) = g(x), & x \in \mathbb{R} \quad (\text{initial velocity}) \end{cases}$$

Factor the wave operator into two directional transport derivatives:

$$\left( \frac{\partial}{\partial t} - c \frac{\partial}{\partial x} \right) \left( \frac{\partial}{\partial t} + c \frac{\partial}{\partial x} \right) u = 0$$

Let $v(x, t) = u_t + c u_x$. Then $v$ satisfies the first-order homogeneous advection equation:

$$v_t - c v_x = 0$$

whose general solution is $v(x, t) = h(x + ct)$ for an arbitrary function $h$. Now solve the inhomogeneous first-order PDE for $u$:

$$u_t + c u_x = h(x + ct)$$

Using the characteristic coordinates $\xi = x - ct$ and $\eta = x + ct$:

$$\frac{\partial^2 u}{\partial \xi \partial \eta} = 0$$

Integrating twice with respect to $\xi$ and $\eta$ yields the general traveling wave solution:

$$u(x, t) = \phi(x - ct) + \psi(x + ct)$$

where $\phi(x - ct)$ represents a right-moving wave at speed $c$, and $\psi(x + ct)$ represents a left-moving wave at speed $c$.

Determining $\phi$ and $\psi$ from Initial Data:

At $t = 0$:

  1. $u(x, 0) = \phi(x) + \psi(x) = f(x)$
  2. $u_t(x, 0) = -c \phi'(x) + c \psi'(x) = g(x)$

Dividing equation (2) by $c$ and integrating from an arbitrary base point $x_0$ to $x$:

$$-\phi(x) + \psi(x) = \frac{1}{c} \int_{x_0}^x g(s)\,ds + K$$

We now have a linear system for $\phi(x)$ and $\psi(x)$:

  • Adding the two equations:
$$2\psi(x) = f(x) + \frac{1}{c}\int_{x_0}^x g(s)\,ds + K \implies \psi(x) = \frac{1}{2}f(x) + \frac{1}{2c}\int_{x_0}^x g(s)\,ds + \frac{K}{2}$$
  • Subtracting the two equations:
$$2\phi(x) = f(x) - \frac{1}{c}\int_{x_0}^x g(s)\,ds - K \implies \phi(x) = \frac{1}{2}f(x) - \frac{1}{2c}\int_{x_0}^x g(s)\,ds - \frac{K}{2}$$

Substituting $\phi(x - ct)$ and $\psi(x + ct)$ into $u(x, t) = \phi(x - ct) + \psi(x + ct)$:

$$u(x, t) = \frac{1}{2}[f(x - ct) + f(x + ct)] + \frac{1}{2c}\left[ \int_{x_0}^{x+ct} g(s)\,ds - \int_{x_0}^{x-ct} g(s)\,ds \right]$$

Combining the integral limits:

Theorem 4.1 (D'Alembert's Formula): The unique solution to the Cauchy problem for the 1D wave equation is:

$$u(x, t) = \frac{1}{2}\left[ f(x - ct) + f(x + ct) \right] + \frac{1}{2c} \int_{x - ct}^{x + ct} g(s)\,ds$$

§4.2 Domain of Dependence, Range of Influence & Strict Causality

1. Spacetime Geometry of Hyperbolic Waves

D'Alembert's formula reveals profound relativistic and geometric properties that distinguish hyperbolic PDEs from parabolic and elliptic equations.

Definition 4.1 (Domain of Dependence): For any spacetime point $(x_0, t_0)$ with $t_0 > 0$, the value $u(x_0, t_0)$ depends exclusively on:

  1. The values of initial displacement $f$ at the two endpoints: $x_0 - c t_0$ and $x_0 + c t_0$.
  2. The values of initial velocity $g$ on the closed spatial interval:
$$D(x_0, t_0) = [x_0 - c t_0, x_0 + c t_0]$$

The interval $D(x_0, t_0)$ is called the domain of dependence of the point $(x_0, t_0)$. In the $(x, t)$ spacetime plane, the triangular region with base $D(x_0, t_0)$ and apex $(x_0, t_0)$ bounded by the characteristic lines $x - ct = x_0 - ct_0$ and $x + ct = x_0 + ct_0$ is the past light cone.


2. Range of Influence

Definition 4.2 (Range of Influence): Conversely, an initial disturbance originating at a point $x = \xi$ at $t = 0$ can only affect spacetime points $(x, t)$ satisfying:

$$\xi - ct \le x \le \xi + ct \iff |x - \xi| \le ct$$

The region $I(\xi) = \{ (x, t) \in \mathbb{R} \times [0, \infty) : |x - \xi| \le ct \}$ is the future light cone or range of influence of the point $\xi$.


3. Strict Causality vs Parabolic Infinite Speed

1. Finite Speed of Propagation: Disturbances propagate at the exact finite speed $c$. If initial data $f$ and $g$ are supported on a compact interval $[a, b]$, then at time $t$, $u(x, t) = 0$ everywhere outside $[a - ct, b + ct]$.

2. Contrast with Diffusion: In the heat equation $u_t = \alpha u_{xx}$, an initial point disturbance at $x = 0$ is instantly felt across the entire universe ($u(x, t) > 0$ for all $x \in \mathbb{R}$ for any $t > 0$, no matter how small). Hyperbolic physics strictly respects relativistic causality.

§4.3 Semi-Infinite & Finite Strings: The Method of Images

1. The Semi-Infinite String with Fixed End (Dirichlet Boundary)

Consider the semi-infinite string $x \ge 0, t \ge 0$ with a clamped boundary at $x = 0$:

$$\begin{cases} u_{tt} - c^2 u_{xx} = 0, & x > 0, \quad t > 0 \\ u(x, 0) = f(x), \quad u_t(x, 0) = g(x), & x > 0 \\ u(0, t) = 0, & t \ge 0 \quad (\text{fixed end}) \end{cases}$$

where $f(0) = g(0) = 0$ for compatibility.

For points where $x \ge ct$, the backward characteristic $x - ct \ge 0$ does not touch the boundary $x = 0$, so D'Alembert's formula holds directly. For points where $x < ct$, the characteristic $x - ct < 0$ reflects off the wall $x = 0$.

The Method of Odd Reflection:

To enforce $u(0, t) = 0$ automatically, extend the initial data to the entire real line $\mathbb{R}$ as odd functions:

$$f_{\text{odd}}(x) = \begin{cases} f(x), & x > 0 \\ 0, & x = 0 \\ -f(-x), & x < 0 \end{cases}, \qquad g_{\text{odd}}(x) = \begin{cases} g(x), & x > 0 \\ 0, & x = 0 \\ -g(-x), & x < 0 \end{cases}$$

Applying D'Alembert's formula with these odd extensions: For $x < ct$, since $x - ct < 0$:

$$f_{\text{odd}}(x - ct) = -f(-(x - ct)) = -f(ct - x)$$

Thus, the solution for $0 < x < ct$ is:

$$u(x, t) = \frac{1}{2}[f(x + ct) - f(ct - x)] + \frac{1}{2c}\int_{ct - x}^{x + ct} g(s)\,ds$$

Physical Interpretation: The term $-f(ct - x)$ represents an inverted wave packet reflected from the fixed boundary with a phase shift of $\pi$ (inversion).


2. Free End (Neumann Boundary Condition)

If the end at $x = 0$ is free to slide frictionlessly on a vertical rod, the boundary condition is $u_x(0, t) = 0$. By applying even extensions $f_{\text{even}}(-x) = f(x)$, the reflected wave maintains its sign:

$$u(x, t) = \frac{1}{2}[f(x + ct) + f(ct - x)] + \frac{1}{2c}\left[ \int_0^{ct - x} g(s)\,ds + \int_0^{x + ct} g(s)\,ds \right]$$

No inversion occurs upon reflection from a free boundary.

§4.4 Separation of Variables: Normal Modes & Standing Waves

1. Separation of Variables on a Finite Interval $[0, L]$

Consider the vibrating string of finite length $L$ with clamped endpoints:

$$\begin{cases} u_{tt} = c^2 u_{xx}, & 0 < x < L, \quad t > 0 \\ u(0, t) = 0, \quad u(L, t) = 0, & t \ge 0 \\ u(x, 0) = f(x), \quad u_t(x, 0) = g(x), & 0 \le x \le L \end{cases}$$

Assume a product solution of the form:

$$u(x, t) = X(x) T(t)$$

Substituting into the wave equation:

$$X(x) T''(t) = c^2 X''(x) T(t) \implies \frac{T''(t)}{c^2 T(t)} = \frac{X''(x)}{X(x)} = -\lambda$$

where $-\lambda$ is a separation constant.

Spatial Boundary Value Problem (Sturm-Liouville Eigenvalue Problem):
$$X''(x) + \lambda X(x) = 0, \qquad X(0) = 0, \quad X(L) = 0$$
  • If $\lambda \le 0$, only trivial solutions $X \equiv 0$ exist.
  • For $\lambda > 0$, set $\lambda = k^2$:
$$X(x) = A \cos(kx) + B \sin(kx)$$

$X(0) = 0 \implies A = 0$. $X(L) = B \sin(kL) = 0 \implies k_n L = n\pi \implies k_n = \frac{n\pi}{L}, \quad n = 1, 2, 3, \dots$ Thus, the eigenvalues and eigenfunctions are:

$$\lambda_n = \left(\frac{n\pi}{L}\right)^2, \qquad X_n(x) = \sin\left(\frac{n\pi x}{L}\right), \quad n \in \mathbb{N}$$
Temporal Equation:
$$T_n''(t) + c^2 k_n^2 T_n(t) = 0 \implies T_n''(t) + \omega_n^2 T_n(t) = 0$$

where the natural circular frequencies are:

$$\omega_n = c k_n = \frac{n\pi c}{L}$$

The general solution for $T_n(t)$ is:

$$T_n(t) = A_n \cos(\omega_n t) + B_n \sin(\omega_n t)$$

2. General Fourier Superposition

By the principle of linear superposition, the general solution is:

$$u(x, t) = \sum_{n=1}^\infty \left[ A_n \cos\left(\frac{n\pi c t}{L}\right) + B_n \sin\left(\frac{n\pi c t}{L}\right) \right] \sin\left(\frac{n\pi x}{L}\right)$$

The Fourier coefficients are determined from initial data by orthogonality:

$$A_n = \frac{2}{L} \int_0^L f(x) \sin\left(\frac{n\pi x}{L}\right)\,dx$$
$$B_n = \frac{2}{n\pi c} \int_0^L g(x) \sin\left(\frac{n\pi x}{L}\right)\,dx$$

§4.5 Uniqueness and Stability: Conservation of Energy & Energy Integrals

1. Total Mechanical Energy of the Vibrating String

The kinetic energy of the string of linear mass density $\rho$ is:

$$E_K(t) = \frac{1}{2} \int_0^L \rho \left(\frac{\partial u}{\partial t}\right)^2\,dx$$

The potential (elastic strain) energy under tension $T$ is:

$$E_P(t) = \frac{1}{2} \int_0^L T \left(\frac{\partial u}{\partial x}\right)^2\,dx$$

Dividing by $\rho$ (recalling $c^2 = T/\rho$), the normalized total energy is defined as:

$$E(t) = \frac{1}{2} \int_0^L \left[ u_t^2 + c^2 u_x^2 \right]\,dx$$

2. Rigorous Proof of Conservation of Energy

Theorem 4.2 (Conservation of Energy): Let $u(x, t)$ be a $C^2$ solution of the wave equation $u_{tt} = c^2 u_{xx}$ on $[0, L] \times [0, \infty)$ satisfying either Dirichlet boundaries ($u(0, t) = u(L, t) = 0$) or Neumann boundaries ($u_x(0, t) = u_x(L, t) = 0$). Then the total energy is strictly constant in time:

$$\frac{dE}{dt} = 0 \implies E(t) = E(0) \quad \forall t \ge 0$$

Proof: Differentiate $E(t)$ under the integral sign:

$$\frac{dE}{dt} = \frac{1}{2} \int_0^L \frac{\partial}{\partial t} \left[ u_t^2 + c^2 u_x^2 \right]\,dx = \int_0^L \left[ u_t u_{tt} + c^2 u_x u_{xt} \right]\,dx$$

Since $u$ is $C^2$, by Schwarz's theorem $u_{xt} = u_{tx}$. Apply integration by parts to the second term:

$$\int_0^L c^2 u_x u_{tx}\,dx = \left[ c^2 u_x u_t \right]_0^L - \int_0^L c^2 u_{xx} u_t\,dx$$

Substitute this back into the derivative of energy:

$$\frac{dE}{dt} = \int_0^L u_t \left[ u_{tt} - c^2 u_{xx} \right]\,dx + \left[ c^2 u_x u_t \right]_0^L$$
  1. In the interior, $u_{tt} - c^2 u_{xx} = 0$ by the wave equation.
  2. At the boundaries $x = 0$ and $x = L$:
  • For Dirichlet conditions, $u(0, t) = u(L, t) = 0 \implies u_t(0, t) = u_t(L, t) = 0$.
  • For Neumann conditions, $u_x(0, t) = u_x(L, t) = 0$.

In either case, the boundary term vanishes identically: $[c^2 u_x u_t]_0^L = 0$. Therefore:

$$\frac{dE}{dt} = 0 \implies E(t) = E(0) \quad \forall t \ge 0$$

$\blacksquare$


3. Proof of Solution Uniqueness via the Energy Method

Theorem 4.3 (Uniqueness of Solutions to the Wave Equation): The initial-boundary value problem for the 1D wave equation with Dirichlet boundary conditions has at most one $C^2$ solution.

Proof: Suppose $u_1$ and $u_2$ are two $C^2$ solutions with identical initial conditions and boundary conditions. Define the difference function $w(x, t) = u_1(x, t) - u_2(x, t)$. By linearity:

  • $w_{tt} - c^2 w_{xx} = 0$ on $(0, L) \times (0, \infty)$
  • $w(x, 0) = f(x) - f(x) = 0$
  • $w_t(x, 0) = g(x) - g(x) = 0$
  • $w(0, t) = 0$ and $w(L, t) = 0$

Now evaluate the energy of the difference function at $t = 0$:

$$E_w(0) = \frac{1}{2}\int_0^L [w_t(x, 0)^2 + c^2 w_x(x, 0)^2]\,dx = \frac{1}{2}\int_0^L [0 + c^2(0)^2]\,dx = 0$$

By Theorem 4.2 (Conservation of Energy):

$$E_w(t) = E_w(0) = 0 \quad \forall t \ge 0$$

Since the integrand $w_t^2 + c^2 w_x^2 \ge 0$ is non-negative and continuous, $E_w(t) = 0$ forces:

$$w_t(x, t) \equiv 0 \quad \text{and} \quad w_x(x, t) \equiv 0 \quad \forall (x, t) \in [0, L] \times [0, \infty)$$

Since all first derivatives vanish identically on a connected domain, $w(x, t)$ is constant:

$$w(x, t) = C$$

Using the boundary condition $w(0, t) = 0$, we find $C = 0$. Thus:

$$w(x, t) \equiv 0 \implies u_1(x, t) \equiv u_2(x, t)$$

The solution is strictly unique. $\blacksquare$

Rigorous Tiered Solved Examination Problems

Step-by-step unskipped derivations, complete proofs, and verification across Foundational, Advanced, and Honors tiers.

Foundational Example 4.1: D'Alembert Solution with Symmetric Triangular Displacement

Solve the infinite wave equation:

$$u_{tt} = c^2 u_{xx}, \qquad x \in \mathbb{R}, \quad t > 0$$

with initial velocity $g(x) \equiv 0$ and initial displacement given by the symmetric triangular pluck:

$$f(x) = \begin{cases} h \left(1 - \frac{|x|}{a}\right), & |x| \le a \\ 0, & |x| > a \end{cases}$$
  1. Write the explicit piecewise expression for $u(x, t)$ for $t > a/c$.
  2. Sketch and describe the evolution of the two separating wave pulses.

1. Application of D'Alembert's Formula

Since the initial velocity is zero ($g(x) = 0$), D'Alembert's formula simplifies to:

$$u(x, t) = \frac{1}{2} [f(x - ct) + f(x + ct)]$$

This represents the superposition of two identical triangular pulses, each of half the original height $h/2$, propagating in opposite directions at constant velocity $\pm c$.


2. Piecewise Analysis for Large Times ($t > a/c$)

When $t > a/c$, the two pulses have completely separated because their centers are at $x = ct$ and $x = -ct$, and the distance between their centers is $2ct > 2a$ (which exceeds the sum of their half-widths $a + a = 2a$).

1. Right-Traveling Pulse: Supported on $[ct - a, ct + a]$:

$$f(x - ct) = h \left( 1 - \frac{|x - ct|}{a} \right) \quad \text{for } |x - ct| \le a$$

Contribution: $\frac{h}{2} \left( 1 - \frac{|x - ct|}{a} \right)$.

2. Left-Traveling Pulse: Supported on $[-ct - a, -ct + a]$:

$$f(x + ct) = h \left( 1 - \frac{|x + ct|}{a} \right) \quad \text{for } |x + ct| \le a$$

Contribution: $\frac{h}{2} \left( 1 - \frac{|x + ct|}{a} \right)$.

3. Intermediate Region: For $-ct + a < x < ct - a$, both pulses vanish, so:

$$u(x, t) = 0$$

Therefore, for $t > a/c$, the complete solution is:

$$u(x, t) = \begin{cases} \frac{h}{2}\left(1 - \frac{|x + ct|}{a}\right), & -ct - a \le x \le -ct + a \\ 0, & -ct + a < x < ct - a \\ \frac{h}{2}\left(1 - \frac{|x - ct|}{a}\right), & ct - a \le x \le ct + a \\ 0, & |x| > ct + a \end{cases}$$

3. Physical Behavior

At $t = 0$, the single triangular tent of peak height $h$ is at rest. As $t$ advances from $0$ to $a/(2c)$, the peak flattens into a plateau of height $h/2$ as the right and left components slide past each other. At $t = a/c$, the two triangles touch at the origin with zero height. For $t > a/c$, two separate triangular pulses of height $h/2$ and base $2a$ travel indefinitely to $\pm\infty$ without attenuation or dispersion. $\blacksquare$

Advanced Example 4.2: Semi-Infinite String with Fixed End and Incoming Gaussian Pulse

Consider the semi-infinite string $x > 0$ with fixed clamped end $u(0, t) = 0$. The string is initially at rest with initial velocity $g(x) \equiv 0$ and an initial Gaussian displacement centered at $x_0 > 0$:

$$f(x) = A \exp\left( -\frac{(x - x_0)^2}{2\sigma^2} \right)$$

where $\sigma \ll x_0$. Using the method of odd reflection:

  1. Find the exact solution $u(x, t)$ for all $x > 0, t > 0$.
  2. Analyze the reflected wave packet and prove that reflection produces an exact phase inversion.

1. Odd Reflection Extension

To satisfy the Dirichlet boundary condition $u(0, t) = 0$ for all $t \ge 0$, we extend $f(x)$ to the entire real line as an odd function $f_{\text{odd}}(x)$:

$$f_{\text{odd}}(x) = \begin{cases} f(x) = A e^{-(x - x_0)^2 / (2\sigma^2)}, & x > 0 \\ 0, & x = 0 \\ -f(-x) = -A e^{-(-x - x_0)^2 / (2\sigma^2)} = -A e^{-(x + x_0)^2 / (2\sigma^2)}, & x < 0 \end{cases}$$

Notice that the fictitious image source is located at $-x_0$ with negative amplitude $-A$.


2. D'Alembert Representation

Since $g \equiv 0$, the solution on $x > 0$ is:

$$u(x, t) = \frac{1}{2} [ f_{\text{odd}}(x - ct) + f_{\text{odd}}(x + ct) ]$$

Since $x > 0$ and $t > 0$, $x + ct > 0$ always; hence:

$$f_{\text{odd}}(x + ct) = f(x + ct) = A e^{-(x + ct - x_0)^2 / (2\sigma^2)}$$

For the term $x - ct$:

  • If $x \ge ct$ (before reflection reaches point $x$):
$$x - ct \ge 0 \implies f_{\text{odd}}(x - ct) = f(x - ct) = A e^{-(x - ct - x_0)^2 / (2\sigma^2)}$$

So for $x \ge ct$:

$$u(x, t) = \frac{A}{2} \left[ e^{-(x - ct - x_0)^2 / (2\sigma^2)} + e^{-(x + ct - x_0)^2 / (2\sigma^2)} \right]$$
  • If $x < ct$ (after reflection has occurred):
$$x - ct < 0 \implies f_{\text{odd}}(x - ct) = -f(-(x - ct)) = -f(ct - x) = -A e^{-(ct - x - x_0)^2 / (2\sigma^2)} = -A e^{-(x - ct + x_0)^2 / (2\sigma^2)}$$

So for $x < ct$:

$$u(x, t) = \frac{A}{2} \left[ e^{-(x + ct - x_0)^2 / (2\sigma^2)} - e^{-(x - ct + x_0)^2 / (2\sigma^2)} \right]$$

3. Analysis of Phase Inversion

At $t = x_0/c$, the incoming left-traveling pulse $\frac{A}{2}e^{-(x + ct - x_0)^2 / (2\sigma^2)}$ strikes the boundary $x = 0$. For $t > x_0/c$, the reflected pulse travels toward the right in the region $x > 0$, described by:

$$u_{\text{refl}}(x, t) = -\frac{A}{2} \exp\left( -\frac{(x - c(t - x_0/c))^2}{2\sigma^2} \right)$$

1. Negative Amplitude: The amplitude is $-\frac{A}{2}$, which is an exact vertical flip (phase shift of $\pi$).

2. Boundary Verification: At $x = 0$:

$$u(0, t) = \frac{A}{2} \left[ e^{-(ct - x_0)^2 / (2\sigma^2)} - e^{-( -ct + x_0)^2 / (2\sigma^2)} \right]$$

Since $(ct - x_0)^2 = (-ct + x_0)^2$, the two terms cancel identically:

$$u(0, t) \equiv 0 \quad \forall t \ge 0$$

The boundary condition is strictly preserved. $\blacksquare$

Honors Challenge Example 4.3: Conservation of Energy and Dirichlet Wave Uniqueness

Consider the inhomogeneous 1D wave equation:

$$u_{tt} - c^2 u_{xx} = F(x, t), \qquad 0 < x < L, \quad t > 0$$

subject to time-dependent boundary conditions $u(0, t) = h_1(t), u(L, t) = h_2(t)$, and initial conditions $u(x, 0) = f(x), u_t(x, 0) = g(x)$.

  1. Formulate the total mechanical energy integral for the homogeneous problem.
  2. Prove that the initial-boundary value problem has at most one classical $C^2$ solution.
  3. Show that the solution depends continuously on the initial data in the energy norm.

1. Difference Function and Homogeneous System

Suppose $u(x, t)$ and $v(x, t)$ are two $C^2$ solutions satisfying the same inhomogeneous PDE, boundary conditions, and initial data. Define $w(x, t) = u(x, t) - v(x, t)$. Then $w(x, t)$ satisfies:

$$\begin{cases} w_{tt} - c^2 w_{xx} = F(x, t) - F(x, t) = 0, & 0 < x < L, \quad t > 0 \\ w(0, t) = h_1(t) - h_1(t) = 0, & t \ge 0 \\ w(L, t) = h_2(t) - h_2(t) = 0, & t \ge 0 \\ w(x, 0) = f(x) - f(x) = 0, & 0 \le x \le L \\ w_t(x, 0) = g(x) - g(x) = 0, & 0 \le x \le L \end{cases}$$

2. Proof of Energy Conservation for $w$

Define the energy of the error function $w$:

$$E[w](t) = \frac{1}{2} \int_0^L \left[ (w_t(x, t))^2 + c^2 (w_x(x, t))^2 \right]\,dx$$

Taking the time derivative:

$$\frac{dE}{dt} = \int_0^L \left[ w_t w_{tt} + c^2 w_x w_{xt} \right]\,dx$$

Integrate the second term by parts with respect to $x$:

$$\int_0^L c^2 w_x w_{xt}\,dx = \left[ c^2 w_x w_t \right]_0^L - \int_0^L c^2 w_{xx} w_t\,dx$$

Combine the terms:

$$\frac{dE}{dt} = \int_0^L w_t (w_{tt} - c^2 w_{xx})\,dx + c^2 w_x(L, t) w_t(L, t) - c^2 w_x(0, t) w_t(0, t)$$
  • $w_{tt} - c^2 w_{xx} = 0$ in $(0, L)$.
  • Since $w(0, t) = 0$ for all $t$, differentiating with respect to $t$ gives $w_t(0, t) = 0$.
  • Since $w(L, t) = 0$ for all $t$, differentiating with respect to $t$ gives $w_t(L, t) = 0$.

Therefore, both boundary terms vanish:

$$\frac{dE}{dt} = 0 \implies E[w](t) = E[w](0) \quad \forall t \ge 0$$

3. Evaluation of Initial Energy and Uniqueness

At $t = 0$:

$$w_t(x, 0) = 0, \qquad w_x(x, 0) = \frac{d}{dx}[w(x, 0)] = \frac{d}{dx}[0] = 0$$

Hence:

$$E[w](0) = \frac{1}{2}\int_0^L [0^2 + c^2(0)^2]\,dx = 0$$

Consequently:

$$E[w](t) = 0 \quad \forall t \ge 0$$

Because the integrand $w_t^2 + c^2 w_x^2$ is continuous and non-negative:

$$w_t(x, t) = 0 \quad \text{and} \quad w_x(x, t) = 0 \quad \forall (x, t) \in [0, L] \times [0, \infty)$$

This implies $w(x, t) = C$ (constant). Since $w(0, t) = 0$, $C = 0$. Therefore:

$$w(x, t) \equiv 0 \implies u(x, t) \equiv v(x, t)$$

The solution is strictly unique.


4. Continuous Dependence on Initial Data

Let $u$ and $\tilde{u}$ be solutions with initial data $(f, g)$ and $(\tilde{f}, \tilde{g})$. By linearity, their difference $\delta u = u - \tilde{u}$ satisfies:

$$E[\delta u](t) = E[\delta u](0) = \frac{1}{2}\int_0^L \left[ (g - \tilde{g})^2 + c^2 (f' - \tilde{f}')^2 \right]\,dx$$

If $\|g - \tilde{g}\|_{L^2} < \epsilon$ and $\|f' - \tilde{f}'\|_{L^2} < \epsilon$, then:

$$E[\delta u](t) \le \frac{1}{2}(1 + c^2)\epsilon^2$$

By Poincaré's inequality, since $\delta u(0, t) = 0$:

$$\max_{x \in [0, L]} |\delta u(x, t)|^2 \le L \int_0^L |\delta u_x|^2\,dx \le \frac{2L}{c^2} E[\delta u](t) \le \frac{L(1+c^2)}{c^2}\epsilon^2$$

Thus, small variations in initial data result in uniformly small variations in the solution for all future times, establishing Hadamard well-posedness. $\blacksquare$