Defining the Boundary-Value Problem
A standard second-order boundary-value problem involves a differential equation defined over an interval $[a, b]$, where the state of the system is pinned down at both ends. This is mathematically expressed as:
$y^{\prime \prime}=f(x, y, y^{\prime}), \quad \text { for } a \leq x \leq b$
with the Dirichlet boundary conditions:
$y(a)=\alpha \quad \text { and } \quad y(b)=\beta$
Unlike IVPs, which require $y(a)$ and $y'(a)$ at a single point, BVPs specify $y$ at $a$ and $b$. We no longer know the "initial slope" $y'(a)$; instead, we must determine a trajectory that "connects the dots" while satisfying the governing equation throughout the interior.
Existence and Uniqueness (Theorem 11.1)
While the Picard–Lindelöf theorem provides local uniqueness for IVPs, BVPs are governed by global behavior. Even a simple linear ODE may have no solution, one unique solution, or infinitely many solutions depending on the domain length $(b-a)$. A unique solution is guaranteed if:
- $f, f_y, \text{ and } f_{y'}$ are continuous on the domain.
- $f_y > 0$ (This acts like a "restoring force" ensuring the solution doesn't fly off to infinity).
- $|f_{y'}|$ is bounded by a constant $M$.
Real-World Application: Structural Deflection
Consider a structural beam of length $l$ subject to a uniform load $q$ and a horizontal tensile force $S$. The deflection $w(x)$ is governed by:
$\frac{d^2 w}{d x^2}(x)=\frac{S}{E I} w(x)+\frac{q x}{2 E I}(x-l)$
With boundary conditions $w(0)=0$ and $w(l)=0$. Here, the beam's ends are anchored, and we must find the curve $w(x)$ describing the physical shape of the beam under stress.