DIFFERENTIAL EQUATIONS • SERIES AND NUMERICAL METHODS

Power Series Method

Solve differential equations by expressing solutions as infinite polynomials built term by term.

Historical Context & Motivation

Not every differential equation can be solved with neat, closed-form expressions like sines, cosines, or exponentials. Mathematicians discovered centuries ago that many important equations in physics and engineering resist these familiar solutions. The power series method was developed as a powerful alternative: instead of hunting for a known function, you assume the solution is an infinite polynomial and then figure out its coefficients one by one. This idea traces back to some of the greatest minds in the history of mathematics.

1668
Newton's Early Series Work
Isaac Newton develops methods for expanding functions as infinite series, laying the groundwork for representing solutions to equations as power series. His work on the generalized binomial theorem opens the door to infinite polynomial representations.
1715
Taylor's Theorem Published
Brook Taylor formally publishes the result that any smooth function can be written as an infinite sum of powers of x, each multiplied by a coefficient involving the function's derivatives. This Taylor series becomes the backbone of the power series method.
1768
Euler Applies Series to Differential Equations
Leonhard Euler systematically uses series expansions to solve differential equations that arise in mechanics and astronomy, showing that power series solutions are not just theoretical—they produce real, usable answers.
1874
Frobenius Extends the Method
Ferdinand Frobenius generalizes the power series method to handle equations with singular points, creating the Frobenius method. This extension dramatically widens the range of solvable equations.

The central question the power series method addresses is this: when a differential equation has no obvious solution in terms of functions you already know, how can you still find an exact, usable answer? The method's elegant solution is to build the answer piece by piece, coefficient by coefficient, as an infinite series.

Core Principles & Definitions

Before diving into the method itself, you need to understand a few foundational ideas. A power series is an infinite sum of the form a₀ + a₁x + a₂x² + a₃x³ + ⋯, where the coefficients a₀, a₁, a₂, … are constants you need to determine. Think of it as a polynomial that never ends. The power series method works by assuming your unknown solution y(x) has this form, then substituting it into the differential equation to find the coefficients.

1

Power Series

An infinite sum y = Σ aₙxⁿ from n = 0 to ∞. Each term is a coefficient aₙ multiplied by x raised to the nth power. This is the assumed form of the solution.
2

Radius of Convergence

The range of x-values for which the series actually adds up to a finite number. Outside this radius, the series diverges and becomes meaningless.
3

Ordinary Point

A value x = x₀ where the coefficient functions in the differential equation are well-behaved (analytic). The power series method works directly at ordinary points.
4

Recurrence Relation

A formula that expresses each coefficient aₙ in terms of earlier coefficients. Once you find this relation, you can generate as many terms of the solution as you need.
5

Term-by-Term Operations

Within the radius of convergence, you can differentiate and add power series term by term, just like you would with finite polynomials. This property makes the method work.
KEY TAKEAWAY
Think of the power series method like building with LEGO bricks. You don't know what the final structure looks like, but you know it's made of standard pieces (powers of x). You place each brick one at a time, and the differential equation tells you exactly which brick goes where. After placing enough bricks, the structure—your solution—takes clear shape.

Visual Explanation

The diagram below illustrates the core workflow of the power series method. You start by assuming a general series form for y(x), compute its derivatives, substitute everything into the differential equation, and then match coefficients of like powers of x to find a recurrence relation. That relation lets you calculate each coefficient in terms of the ones before it.

The six-step workflow of the power series method. You assume a general series, differentiate term by term, substitute into the ODE, group like powers of x, set each coefficient grouping equal to zero to obtain a recurrence relation, and finally build the solution from two linearly independent series.

Notice that the final solution typically splits into two parts: one multiplied by the free constant a₀ and another multiplied by a₁. These correspond to the two linearly independent solutions you expect from a second-order ODE. The values of a₀ and a₁ are determined by initial conditions if they are given.

Mathematical Framework

Let's formalize the method. Consider a second-order linear ODE of the form shown below. The power series method applies whenever x = 0 is an ordinary point of the equation, meaning the coefficient functions P(x) and Q(x) can themselves be expressed as power series around x = 0.

STANDARD FORM OF THE ODE
y'' + P(x)·y' + Q(x)·y = 0
y'' is the second derivative of y with respect to x; P(x) and Q(x) are coefficient functions that must be analytic (expressible as power series) at the point of expansion.
ASSUMED SERIES SOLUTION
y = Σ (n=0 to ∞) aₙxⁿ = a₀ + a₁x + a₂x² + a₃x³ + ⋯
The coefficients a₀, a₁, a₂, … are unknown constants. The goal of the method is to determine these coefficients by substituting this series into the ODE.
DERIVATIVES OF THE SERIES
y' = Σ (n=1 to ∞) n·aₙxⁿ⁻¹, y'' = Σ (n=2 to ∞) n(n−1)·aₙxⁿ⁻²
Differentiation is performed term by term. Notice that the starting index shifts: y' starts at n = 1 (since the constant term vanishes), and y'' starts at n = 2.
RECURRENCE RELATION (GENERAL IDEA)
aₙ₊₂ = f(aₙ, aₙ₊₁, n)
After substituting the series into the ODE and equating coefficients of each power of x to zero, you obtain a formula that expresses higher-index coefficients in terms of earlier ones. This recurrence relation is the key output of the method.

The crucial algebraic step is re-indexing. When you substitute y, y', and y'' into the ODE, the summations start at different values of n and the powers of x don't initially match. You re-index (shift the dummy variable) so that every sum is expressed in terms of the same power xⁿ. Only then can you factor out xⁿ and set the combined coefficient equal to zero.

The Re-Indexing Technique in Detail

Re-indexing is the step where many students first feel confused, so let's break it down carefully. When you have y'' = Σ n(n−1)aₙxⁿ⁻² starting at n = 2, the exponent on x is n − 2, not n. To align this with terms like aₙxⁿ in the original y, you introduce a new index. Let m = n − 2, so n = m + 2, and when n = 2 the new index m = 0. The sum becomes Σ (m+2)(m+1)a_{m+2}xᵐ starting at m = 0. Since m is just a label, you can rename it back to n.

Before re-indexing, y, y', and y'' contain different powers of x (xⁿ, xⁿ⁻¹, xⁿ⁻²). After re-indexing, all sums share the common factor xⁿ, allowing you to combine them and set the overall coefficient of each xⁿ equal to zero.

Once every series runs over the same power xⁿ, you can combine them under a single summation sign. Since the sum equals zero for all x, the coefficient of every individual power xⁿ must be zero. This yields the recurrence relation—a formula that tells you each later coefficient in terms of earlier ones. Typically, a₀ and a₁ remain as free constants, reflecting the two arbitrary constants expected in the general solution of a second-order ODE.

Worked Example

Let's solve the equation y'' − y = 0 using the power series method. You might already know the answer (it's eˣ and e⁻ˣ), but walking through the series approach shows exactly how the method works.

Solve y'' − y = 0 by Power Series
1
Step 1 — Assume a Power Series SolutionAssume y = Σ aₙxⁿ (from n = 0 to ∞). Then y' = Σ n·aₙxⁿ⁻¹ (from n = 1) and y'' = Σ n(n−1)·aₙxⁿ⁻² (from n = 2).
2
Step 2 — Substitute into the ODESubstituting into y'' − y = 0 gives: Σ n(n−1)aₙxⁿ⁻² − Σ aₙxⁿ = 0.
3
Step 3 — Re-Index the y'' SumIn the first sum, let m = n − 2 (so n = m + 2). When n = 2, m = 0. The sum becomes Σ (m+2)(m+1)a_{m+2}xᵐ, starting at m = 0. Rename m back to n: Σ (n+2)(n+1)aₙ₊₂xⁿ, from n = 0.
4
Step 4 — Combine and Set Coefficients to ZeroBoth sums now run from n = 0 with the factor xⁿ. Combining: Σ [(n+2)(n+1)aₙ₊₂ − aₙ]xⁿ = 0. Since this holds for all x, each bracket must be zero.
Recurrence relation: aₙ₊₂ = aₙ / [(n+2)(n+1)]
5
Step 5 — Compute CoefficientsStarting from a₀ and a₁ (both free): a₂ = a₀/2!, a₃ = a₁/3!, a₄ = a₀/4!, a₅ = a₁/5!, and so on. The even-indexed coefficients depend on a₀ and the odd-indexed ones depend on a₁.
6
Step 6 — Write the General SolutionGroup even and odd terms: y = a₀(1 + x²/2! + x⁴/4! + ⋯) + a₁(x + x³/3! + x⁵/5! + ⋯). You may recognize these as the Taylor series for cosh(x) and sinh(x), or equivalently, the general solution is y = C₁eˣ + C₂e⁻ˣ.
y = a₀ cosh(x) + a₁ sinh(x) = C₁eˣ + C₂e⁻ˣ

Strengths, Limitations & Comparisons

The power series method is powerful, but it isn't the right tool for every equation. Understanding when to use it—and when to choose an alternative—is just as important as knowing how to apply it.

Strengths and limitations of the power series method
FeatureStrengthLimitation
Range of applicabilityWorks on equations where standard formulas fail, including variable-coefficient ODEsOnly works at ordinary points; singular points require the Frobenius method
ExactnessProduces the exact solution as an infinite series, not an approximationFor practical computation you must truncate the series, introducing truncation error
Ease of useSystematic, algorithmic—follow the same steps every timeAlgebra can become very tedious for complicated coefficient functions
Closed-form recognitionSometimes the series can be identified as a known function (eˣ, sin x, etc.)Often the series does not simplify to a familiar function
🔧 WHEN TO USE WHAT
Think of the power series method like a Swiss army knife—it handles many situations, but a specialized tool may be faster when one exists. If the ODE has constant coefficients, the characteristic equation method is quicker. If the equation is nonlinear or lacks an ordinary point, you'll need other approaches. But for linear, variable-coefficient equations around an ordinary point, the power series method is your best friend.

Connection to Advanced Theory

The power series method is a gateway to several advanced topics in differential equations and mathematical physics. Once you're comfortable with the basic procedure, you're well prepared to study extensions that handle more difficult situations.

How the power series method connects to advanced topics
Power Series MethodAdvanced Extension
Assumes y = Σ aₙxⁿ (integer powers only)Frobenius method: assumes y = xʳ Σ aₙxⁿ, allowing fractional or negative leading exponents near singular points
Works at ordinary points of the ODERegular singular points: the Frobenius method classifies and handles these, producing solutions like Bessel functions and Legendre polynomials
Series may converge only within a finite radiusAnalytic continuation: extends the solution beyond the original radius of convergence into the complex plane
Produces exact series solutionsNumerical methods (Euler, Runge-Kutta): provide approximate solutions when neither closed-form nor series solutions are practical

Many of the most famous functions in science—Bessel functions, Legendre polynomials, and Hermite polynomials—are defined by power series solutions to particular differential equations. Mastering the basic power series method is the first step toward understanding these special functions that appear throughout physics and engineering.

Practice Problems

PROBLEM 1CONCEPTUAL
Explain in your own words why we need to re-index the sums when applying the power series method. What would go wrong if we tried to combine the sums without re-indexing?
PROBLEM 2BASIC CALCULATION
Given the recurrence relation aₙ₊₂ = −aₙ / [(n+2)(n+1)] with a₀ = 1 and a₁ = 0, compute the coefficients a₂, a₃, a₄, a₅, and a₆. Then write out the first four non-zero terms of the series solution.
PROBLEM 3INTERMEDIATE
Use the power series method to find the recurrence relation for the ODE y'' + x·y = 0. Start by assuming y = Σ aₙxⁿ, substitute, re-index, and derive the relation between aₙ₊₂ and earlier coefficients.
PROBLEM 4APPLIED
A spring-mass system with a time-varying spring constant is modeled by y'' + t·y = 0 (the Airy equation). Using the recurrence from Problem 3, compute the first five non-zero terms of the solution satisfying y(0) = 1 and y'(0) = 0. Then approximate y(0.5) by substituting t = 0.5 into your truncated series.
PROBLEM 5CRITICAL THINKING
Consider the ODE y'' + (1/x)·y' + y = 0. Explain why the standard power series method (assuming y = Σ aₙxⁿ about x = 0) cannot be applied directly. What feature of the equation causes the difficulty, and what modification to the method would you need?

Power Series Method — Summary

The power series method solves linear differential equations by assuming the unknown function is an infinite polynomial y = Σ aₙxⁿ. You differentiate the series term by term, substitute it into the ODE, and re-index so that all sums share the same power xⁿ. Setting the combined coefficient of each xⁿ equal to zero yields a recurrence relation that generates every coefficient from the two free constants a₀ and a₁, which correspond to the two linearly independent solutions of a second-order equation.

The method works at ordinary points of the ODE, where the coefficient functions are analytic. When the series converges, it gives the exact solution within a radius of convergence. For singular points, the technique extends to the Frobenius method. Mastering power series solutions prepares you for the special functions—Bessel, Legendre, Hermite—that underpin much of advanced physics and engineering.

Varsity Tutors • Differential Equations • Power Series Method