= Boundary value problem
{wiki}
A <boundary value problem> asks for a <differential equation> solution together with conditions on the <boundary> of a specified <domain>. These can prescribe the value, a <normal derivative>, a flux, or a combination. For an <ordinary differential equation> on an <interval>, the conditions are often distributed between its endpoints; a <partial differential equation> uses the spatial boundary and, for evolution, also needs appropriate initial data.
For the <Dirichlet boundary conditions> in
$$
-u''=1\quad(0<x<1),\qquad u(0)=u(1)=0,
$$
twice integrating gives $u(x)=x(1-x)/2$. The <boundary conditions> determine the two constants. By contrast, for $-u''=f$ with derivative-zero conditions at both ends, integrating the equation requires $\int_0^1f=0$ and leaves an additive constant undetermined. This is the <Neumann boundary-value problem> compatibility and normalization issue.
Existence, uniqueness and <continuous dependence on forcing> belong to the whole operator/domain/boundary specification. A formal differential formula alone does not establish that a <boundary value problem> is solvable.<Ordinary Differential Equations and Dynamical Systems by Gerald Teschl>[Chapter 5, boundary-value and Sturm-Liouville problems]{ref}
Back to article page