01 / 21
\u03A3
\( a_n x^n \)

Power Series
Method for ODEs

Learn to solve differential equations by expressing solutions as infinite power series: when standard methods fail, series save the day.

Differential Equations

What Is a Power Series?

A power series centered at \( x = a \) is an infinite sum of the form:

Each \( a_n \) is a coefficient we need to determine. When \( a = 0 \), this becomes a Maclaurin series: \( \sum a_n x^n \). You already know famous examples: \( e^x = \sum \frac{x^n}{n!} \).

Why Use Power Series for ODEs?

✕ Standard Methods Fail
Many ODEs with variable coefficients like \( y'' + xy = 0 \) cannot be solved with characteristic equations or integrating factors.
✓ Series Always Work
Near an ordinary point, power series representations are guaranteed to converge and give a valid solution.

The Big Idea: ODE to Algebra

Differential Equation
\( y'' + P(x)y' + Q(x)y = 0 \)
assume Substitute
\( y = \sum a_n x^n \)
equate Recurrence Relation
\( a_{n+2} = f(a_n) \)

We assume the solution is a power series, substitute it into the ODE, and match coefficients of each power of \( x \) to find a recurrence for \( a_n \).

Ordinary Points and Analyticity

Consider the standard form ODE: \( y'' + P(x)y' + Q(x)y = 0 \). A point \( x = a \) is called an ordinary point if both \( P(x) \) and \( Q(x) \) are analytic (have convergent Taylor series) at \( x = a \).

Ordinary vs Singular Points

Ordinary Point Singular Point
Definition \( P(x), Q(x) \) analytic at \( x = a \) \( P(x) \) or \( Q(x) \) not analytic at \( x = a \)
Method Direct power series substitution Frobenius method (modified series)
Solutions Two independent series solutions May need \( \ln(x) \) terms
Example \( y'' + xy = 0 \) at \( x = 0 \) \( x^2 y'' + xy' + y = 0 \) at \( x = 0 \)

Key Differentiation Identities

If \( y = \sum_{n=0}^{\infty} a_n x^n \), we can differentiate term by term:

1
First derivative: \( y' = \sum_{n=1}^{\infty} n\, a_n\, x^{n-1} \)
2
Second derivative: \( y'' = \sum_{n=2}^{\infty} n(n-1)\, a_n\, x^{n-2} \)
3
Product with \( x \): \( x \cdot y = \sum_{n=0}^{\infty} a_n\, x^{n+1} \) (shifts the index by 1)

Step-by-Step Procedure

1
Verify ordinary point: Write ODE in standard form \( y'' + P(x)y' + Q(x)y = 0 \). Check \( P \) and \( Q \) are analytic at \( x = a \).
2
Assume a series solution: Let \( y = \sum_{n=0}^{\infty} a_n x^n \). Compute \( y' \) and \( y'' \) term by term.
3
Substitute into ODE: Replace \( y, y', y'' \) with their series forms in the equation.
4
Re-index and combine: Shift indices so all sums use the same power \( x^n \). Factor out \( x^n \).
5
Set coefficients to zero: Since the series equals zero for all \( x \), each coefficient of \( x^n \) must be zero. This gives a recurrence relation.
6
Solve the recurrence: Express all \( a_n \) in terms of \( a_0 \) and \( a_1 \) (the two arbitrary constants). Write the general solution.

Example 1: \( y'' - y = 0 \)

Step 1: Already in standard form with \( P(x) = 0, Q(x) = -1 \). The point \( x = 0 \) is ordinary.

Step 2: Let \( y = \sum_{n=0}^{\infty} a_n x^n \), so \( y'' = \sum_{n=2}^{\infty} n(n-1) a_n x^{n-2} \).

Step 3: Substitute: \( \sum_{n=2}^{\infty} n(n-1) a_n x^{n-2} - \sum_{n=0}^{\infty} a_n x^n = 0 \).

Example 1: Building the Solution

The recurrence \( a_{k+2} = \frac{a_k}{(k+2)(k+1)} \) connects even indices and odd indices separately:

Even terms (from \( a_0 \))
\( a_2 = \frac{a_0}{2!} \), \( a_4 = \frac{a_0}{4!} \), \( a_6 = \frac{a_0}{6!} \), ...
Odd terms (from \( a_1 \))
\( a_3 = \frac{a_1}{3!} \), \( a_5 = \frac{a_1}{5!} \), \( a_7 = \frac{a_1}{7!} \), ...

Visualizing Partial Sums

Adding more terms to the series brings us closer to the true solution \( \cosh x \):

x y 1 2 -1 -2 1 2 3 4 N=0 N=1 N=2 cosh x

Example 2: The Airy Equation

Solve: \( y'' - xy = 0 \). This equation arises in quantum mechanics and optics. Standard methods cannot solve it in closed form.

Substitute \( y = \sum a_n x^n \):

Airy Equation: Tracing the Pattern

The recurrence \( a_{k+2} = \frac{a_{k-1}}{(k+2)(k+1)} \) links every third coefficient:

From \( a_0 \)
\( a_3 = \frac{a_0}{2 \cdot 3} \) \( a_6 = \frac{a_0}{2 \cdot 3 \cdot 5 \cdot 6} \) \( a_9 = \frac{a_0}{2 \cdot 3 \cdot 5 \cdot 6 \cdot 8 \cdot 9} \)
From \( a_1 \)
\( a_4 = \frac{a_1}{3 \cdot 4} \) \( a_7 = \frac{a_1}{3 \cdot 4 \cdot 6 \cdot 7} \) \( a_{10} = \frac{a_1}{3 \cdot 4 \cdot 6 \cdot 7 \cdot 9 \cdot 10} \)
From \( a_2 = 0 \)
\( a_5 = 0 \) \( a_8 = 0 \) \( a_{11} = 0 \) All zero!

Example 3: \( y'' + x^2 y = 0 \)

This is harder because \( x^2 \) shifts the index by two instead of one. Substitute \( y = \sum a_n x^n \):

k=0
\( 2 \cdot 1 \cdot a_2 = 0 \implies \) \( a_2 = 0 \)
k=1
\( 3 \cdot 2 \cdot a_3 = 0 \implies \) \( a_3 = 0 \)
k≥2
\( a_{k+2} = \dfrac{-a_{k-2}}{(k+2)(k+1)} \): links indices 4 apart

Radius of Convergence

The power series solution converges at least as far as the nearest singularity of the ODE:

\( y'' - y = 0 \)
No singular points. \( R = \infty \).
\( (1-x^2)y'' - 2xy' = 0 \)
Singular at \( x = \pm 1 \). Series about \( x = 0 \) has \( R = 1 \).

Common Mistakes to Avoid

✕ Wrong Re-indexing
Forgetting to change the starting index when substituting \( k = n-2 \). If \( n \) starts at 2, then \( k \) starts at 0, not 2.
✕ Ignoring Low-Order Terms
Sums may not start at the same index. The \( k = 0 \) and \( k = 1 \) terms often must be handled separately before applying the general recurrence.
✕ Forgetting Two Constants
A second-order ODE needs two arbitrary constants (\( a_0 \) and \( a_1 \)). If your solution has only one, you are missing a family of solutions.

Question 1 of 5

If \( y = \sum_{n=0}^{\infty} a_n x^n \), what is the series for \( y'' \)?

Question 2 of 5

For the ODE \( y'' - xy = 0 \), the recurrence relation for the coefficients is:

Question 3 of 5

For the ODE \( (1 - x^2)y'' - 2xy' + 6y = 0 \), what is the radius of convergence of a power series solution centered at \( x = 0 \)?

Question 4 of 5

When applying the power series method to a second-order ODE, how many arbitrary constants should the general solution contain?

Question 5 of 5

Which of the following is NOT a valid first step in the power series method?