Differential Equations · Undergraduate
Boundary Value Problems and Eigenvalues
Quick answer
A boundary value problem fixes the solution at two different points, such as y(0) = 0 and y(L) = 0, instead of giving y and y′ at one point. Unlike initial value problems, these can have one solution, none, or infinitely many. For y″ + λy = 0 with y(0) = y(L) = 0, nonzero solutions exist only for the eigenvalues λₙ = (nπ/L)², with eigenfunctions sin(nπx/L). These are the building blocks of Fourier series and the heat and wave equations.
What you'll learn
- Solve a boundary value problem for y″ + k²y = 0
- Recognize when a boundary value problem has no solution or many
- Find the eigenvalues of y″ + λy = 0 with y(0) = y(L) = 0
- Match each eigenvalue with its eigenfunction
Conditions at both ends
A boundary value problem asks for a solution with conditions at two points, such as
The zero function always works. A number for which a nonzero solution exists is an eigenvalue, and that solution is an eigenfunction. For this problem:
- sin x, λ = 1
- sin 2x, λ = 4
- sin 3x, λ = 9
Why only special λ work
For , the general solution is . The condition forces , leaving . Then allows only when , that is , so . For the solutions are lines or exponentials, and the two conditions force them to be .
Worked examples
Common mistakes
Practice problems
-
Find the smallest eigenvalue of , .
Answer
Full solution
.
-
Find the third eigenvalue for the same problem.
Answer
Full solution
.
-
Find the smallest eigenvalue of , .
Answer
Full solution
.
-
Find the second eigenvalue of , .
Answer
Full solution
.
-
Solve with and , then find .
Answer
Full solution
with , so and .
-
Does , , have a solution?
Answer
No
Full solution
Every solution with is , which is at .
-
For , , the eigenvalue has eigenfunction . Find .
Answer
Full solution
.
-
A student says is an eigenvalue of , . How many nonzero solutions does give?
Answer
Full solution
gives ; forces and forces .
Frequently asked questions
How is a boundary value problem different from an initial value problem?
An initial value problem gives y and y′ at one point; a boundary value problem gives conditions at two points.
Can a boundary value problem have no solution?
Yes. y″ + y = 0 with y(0) = 0 and y(π) = 1 has none, since every solution through y(0) = 0 is c sin x, which is 0 at π.
What are the eigenvalues of y″ + λy = 0, y(0) = y(L) = 0?
λₙ = (nπ/L)² for n = 1, 2, 3, …, with eigenfunctions sin(nπx/L).
Is λ = 0 an eigenvalue?
No. Then y = ax + b, and both boundary conditions force a = b = 0.