"

Series Solutions of Differential Equations

5.2 Series Solution to Differential Equations

Power Series Solutions to Linear Differential Equations

In earlier discussions, we primarily focused on homogeneous linear differential equations with constant coefficients. However, many physical applications lead to more complex second-order homogeneous linear differential equations of the form

 P0(x)y+P1(x)y+P2(x)y=0 (5.2.1)

where P0, P1, and P2 are polynomials with no common factor. Often, the solutions to Equation 5.2.1 cannot be expressed in terms of familiar functions, prompting the use of series solutions. We start by normalizing the equation, dividing by P0(x) to make the coefficient of y one.

 y+P1(x)P0(x)y+P2(x)P0(x)y=0

Given the continuity of polynomials, P1/P0, and P2/P0 are continuous except possibly where P0(x0)=0. A point x0 where P0(x0)0 is called an ordinary point of Equation 5.2.1; otherwise it is a singular point. Importantly, at ordinary points, P1/P0  and P2/P0 are analytic, allowing for power series representation.

Theorem. Suppose P0, P1, and P2 are polynomials with no common factor and P0(x)0. If x0  is an ordinary point of Equation 5.2.1, then every solution of the equation can be represented by a power series

 y(x)=n=0 an(x-x0)n (5.2.2)

Moreover, the radius of convergence R of such a power series solution is at least as large as the distance from x0 to the nearest singular point (real or complex) of the equation. If P0 is constant, implying it is never zero, the radius of convergence will be infinity and the interval of convergence will be (-,+).

To find series solutions of Equation 5.2.1, we consider a power series converging near an ordinary point x0. We assume that the solution can be written as a power series 5.2.2, substitute y and its derivatives in the given differential equation, and collect like powers of x-x0. Setting the coefficient of each power to zero, we can systematically solve for the an coefficients, often resulting in a recursive relation.

 

How to Find a Series Solution to a Differential Equation

1. Determine the differential equation and choose the point x0 around which to expand the series (typically an ordinary point)

2. Assume a power series solution (Equation 5.2.2) for y and find its derivatives y , y, etc., as required by the differential equation.

3. Substitute the series and its derivative into the differential equation.

4. Organize like powers of x-x0 by aligning terms, ensuring all series are expressed from the same starting value of n.

5. Collect and group the coefficients of like powers of x-x0.

6. Solve equations by equating coefficients of like powers of x-x0 to find relations among an‘s.

7. Use the given initial or boundary conditions to find specific an‘s. Use the recursive relation to determine all coefficients.

8. Construct the solution with the coefficients found and discuss the radius and interval of convergence.

 

Example 5.2.1: Find a Series Solution to an Equation with Constant Coefficients

Determine a series solution for the differential equation

y+y=0

Show/Hide Solution

 

1. Notice that P1(x)=1 and thus the coefficients are analytic at every point. We assume x0=0 and that the solution can be written as a power series

  y=n=0 anxn

 

2. First, we need to find y:

 y=ddxn=0 anxn =n=1 nanxn-1

 y=ddxn=1 nanxn-1 =n=2 n(n-1)anxn-2

3. Next, we substitute y and y into the equation:

y+y=0

 n=2 n(n-1)anxn-2 +n=0 anxn=0

4. The next step is to align terms. To do this we need to shift the summation indices to start at the same value. Letting k=n-2 or equally n=k+2 in the first summation and n=k in the second summation, we have

 k=0 (k+2)(k+1)ak+2xk +k=0 akxk=0

5. Adding the series yields

 k=0 [(k+2)(k+1)ak+2+ak]xk=0

6. From the Power Series Vanishing on an Interval property discussed in Section 5.1, we know that If a power series is zero for all x, then all its coefficients must be zero. Therefore, we conclude that

 (k+2)(k+1)ak+2+ak=0

or

 ak+2=-ak(k+2)(k+1),   k0

This is called the recurrence relation for the values of k for which the relation is true.

7. Next, we write a few terms of the series to see if we can determine the trend and hopefully the explicit formula of the series. Setting k=0,1,2,...,5, we get

 k=0  a2=-a0(2)(1)  k=1  a3=-a1(3)(2)
 k=2  a4=-a2(4)(3) =a0(4)(3)(2)(1)  k=3  a5=-a3(5)(4) =a1(5)(4)(3)(2)
 k=4  a6=-a4(6)(5) =-a0(6)(5)(4)(3)(2)(1)  k=5  a7=-a5(7)(6) =-a1(7)(6)(5)(4)(3)(2)

Notice that the term with even indices can be written in terms of the previous term and eventually in terms of a0 and so can be the odd indices in terms of a1. Therefore, by writing the recurrence relation separately for odd (k=2m+1) and even (k=2m) indices, we get

 a2m=(-1)ma0(2m)!   , m0

 a2m+1=(-1)ma1(2m+1)!   , m0

8. Thus the general solution of the equation can be written as

  y=n=0 anxn

 =m=0 a2mx2m +m=0 a2m+1x2m+1

 =a0m=0 (-1)mx2m(2m)! +a1m=0 (-1)mx2m+1(2m+1)!

We recognize that the series in the solution are the Maclaurin series of cos(x) and sin(x), respectively.

 cos(x)=m=0 (-1)mx2m(2m)!    and   sin(x)=m=0 (-1)mx2m+1(2m+1)!

Therefore, the general solution can be expressed as

 y=a0cos(x)+a1sin(x)

for some arbitrary constant a0  and a1. This is the same solution we would obtain using the methods learned in previous sections.

The interval of convergence for both the cosine and sine series is all real numbers (-,).

For both series in the solution, the Ratio Test indicates that as m the limit L approaches zero, which means the series converge for all real numbers. Therefore, without prior knowledge of the series representing sine and cosine, we would conclude that the interval of convergence for each series and hence the combined series solution is all real number (-,).

 

Try an Example

 

In practice, we are interested in finding the series solution for equations with nonconstant coefficients. This is because equations with constant coefficients can be easily solved using the technique outlined in Chapter 3 for homogeneous equations with constant coefficients. Let us do another example for an equation with nonconstant coefficients.

 

Example 5.2.2: Find a Series Solution to an Equation with Variable Coefficents

Find a series solution for the differential equation

(1+x2)y+2xy-2y=0 

Show/Hide Solution

 

1. Note that P1(x)=1+x2 has no root and thus every point for this equation is an ordinary point. We assume that the solution can be written as a power series

  y=n=0 anxn

2. Next, we find y and y :

 y=ddxn=0 anxn =n=1 nanxn-1

 y=ddxn=1 nanxn-1 =n=2 n(n-1)anxn-2

3. Next, we substitute y, y, and y into the equation:

 (1+x2)n=2 n(n-1)anxn-2 +2xn=1 nanxn-1 -2n=0 anxn=0

Multiplying the coefficients by the series, we get

 n=2 n(n-1)anxn-2 +n=2 n(n-1)anxn +2n=1 nanxn -2n=0 anxn=0

4. Note that the exponent of x is the same in all but the first series. Therefore, we only need to shift the index of the first summations by 2:

 n=0 (n+2)(n+1)an+2xn +n=2 n(n-1)anxn +2n=1 nanxn -2n=0 anxn=0

Also, notice that the second series is zero at n=0,1 . So its index can start at n=0 . Likewise, the third series is zero at n=0 so it can too start at n=0 . Then we rewrite the indices and we have

 n=0 (n+2)(n+1)an+2xn +n=0 n(n-1)anxn +2n=0 nanxn -2n=0 anxn=0

5. Combining the series yields

 n=0 [(n+2)(n+1)an+2+(n(n-1)+2n-2)an]xn=0

 n=0 [(n+2)(n+1)an+2+(n2+n-2)an]xn=0

6. Now setting the coefficient to zero gives

 (n+2)(n+1)an+2+(n2+n-2)an=0

 an+2=-n2+n-2(n+2)(n+1)an

 =-(n+2)(n-1)(n+2)(n+1)an

7. So the recursive relation is simplified to

 an+2=-n-1n+1an

Setting n=0,1,2,...,5 , we get

 n=0  a2=a0  n=1  a3=0a1=0
 n=2  a4=-13a2 =-13a0  n=3  a5=-24a3 =0
 n=4  a6=-35a4 =15a0  n=5  a7=-46a5 =0

Notice that all the terms with odd indices are zero except a1. Therefore, by writing the recurrence relation separately for odd (n=2m+1) and even (n=2m) indices, we obtain

 a2m=(-1)m+112m-1a0   , m1

 a2m+1=a1   , m=0

8. Thus the general solution of the equation can be written as

  y=n=0 anxn

 =a0+a1x2(0)+1+m=1 a2mx2m

 =a0+a1x+a0m=1(-1)m+1x2m2m-1

 

Example 5.2.3: Find a Series Solution to an Equation with Variable Coefficents

Find the first six terms in the series solution of the initial value problem

(1+x2)y+2xy-2y=0,  y(0)=2, y(0)=3 

Show/Hide Solution

 

In Example 5.2.2, we found the general series solution to this differential equation.

 =a0+a1x+a0m=1(-1)m+1x2m2m-1

To apply the initial conditions, we first recognize that a0=y(0)=2 and a1=y(0)=3. Then, we substitute a0  and a1 into the general solution to compute the other terms.

a0=2

a1=3

 m=1   a2= a0[(-1)1+1x2(1)2(1)-1]   a2=2x2

a3=0

 m=2   a4=a0[(-1)2+1x2(2)2(2)-1]   a4=-23x4

a5=0

 m=3   a6=a0[(-1)3+1x2(3)2(3)-1]   a6=25x6

a7=0

 m=4   a8=a0[(-1)4+1x2(4)2(4)-1]   a8=-27x8

Therefore, the solution to the initial value problem is

 y(x)=2+3x+2x2-23x4+25x6-27x8+... .

 

Try an Example

 

Section 5.2 Exercises

  1. Find the first five terms in the series solution of the initial value problem

    (1+x)y+(1-3x)y-y=0,  y(0)=3, y(0)=2 

    Show/Hide Answer

     y=3+2x+12x2+x3-1124x4+...

     

  2. Find the first five terms in the series solution of the initial value problem

    y-2xy+y=0,  y(0)=4, y(0)=3 

    Show/Hide Answer

    y=4+3x-2x2+12x3-12x4+...

     

  3. Find the first five terms in the series solution of the initial value problem

    (2+x)y+(1-4x)y+(2+5x)y=0,  y(0)=1, y(0)=2 

    Show/Hide Answer

     y=1+2x-x2+14x3-7396x4+...

License

Icon for the Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License

Differential Equations Copyright © 2024 by Amir Tavangar is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License, except where otherwise noted.