We now investigate the solution of the homogeneous second-order linear equation
near a singular point. Recall that if the functions A, B, and C are polynomials having no common factors, then the singular points of Eq. (1) are simply those points where A(x)=0.
whereas the Legendre equation of order n,
has the two singular points x=−1
We will restrict our attention to the case in which x=0
and 1−x2=1−(t+1)2=−2t−t2,
This new equation has the singular point t=0
A differential equation having a singular point at 0 ordinarily will not have power series solutions of the form y(x)=∑cnxn,
where P=B/A
we see that P(x)=1/x
We will see presently that the power series method can be generalized to apply near the singular point x=0
where
In particular, the singular point x=0
noting that p(x)≡1
By contrast, consider the equation
which has the singular point x=0.
Because
as x→0
Consider the differential equation
In the standard form y″+Py′+Qy=0
Because
both approach ∞
Thus
Because a quotient of polynomials is analytic wherever the denominator is nonzero, we see that p(x) and q(x) are both analytic at x=0.
It may happen that when we begin with a differential equation in the general form in Eq. (1) and rewrite it in the form in (3), the functions p(x) and q(x) as given in (4) are indeterminate forms at x=0.
and
If p0=0=q0, then x=0 may be an ordinary point of the differential equation x2y″+xp(x)y′+q(x)y=0 in (3). Otherwise:
If both the limits in (5) and (6) exist and are finite, then x=0 is a regular singular point.
If either limit fails to exist or is infinite, then x=0 is an irregular singular point.
The most common case in applications, for the differential equation written in the form
is that the functions p(x) and q(x) are polynomials. In this case p0=p(0) and q0=q(0) are simply the constant terms of these polynomials, so there is no need to evaluate the limits in Eqs. (5) and (6).
To investigate the nature of the point x=0 for the differential equation
we first write it in the form in (3):
Then l’Hôpital’s rule gives the values
and
for the limits in (5) and (6). Since they are not both zero, we see that x=0 is not an ordinary point. But both limits are finite, so the singular point x=0 is regular. Alternatively, we could write
and
These (convergent) power series show explicitly that p(x) and q(x) are analytic and moreover that p0=p(0)=1 and q0=q(0)=12, thereby verifying directly that x=0 is a regular singular point.
We now approach the task of actually finding solutions of a second-order linear differential equation near the regular singular point x=0. The simplest such equation is the constant-coefficient equidimensional equation
to which Eq. (3) reduces when p(x)≡p0 and q(x)≡q0 are constants. In this case we can verify by direct substitution that the simple power function y(x)=xr is a solution of Eq. (7) if and only if r is a root of the quadratic equation
In the general case, in which p(x) and q(x) are power series rather than constants, it is a reasonable conjecture that our differential equation might have a solution of the form
—the product of xr and a power series. This turns out to be a very fruitful conjecture; according to Theorem 1 (soon to be stated formally), every equation of the form in (1) having x=0 as a regular singular point does, indeed, have at least one such solution. This fact is the basis for the method of Frobenius, named for the German mathematician Georg Frobenius (1848–1917), who discovered the method in the 1870s.
An infinite series of the form in (9) is called a Frobenius series. Note that a Frobenius series is generally not a power series. For instance, with r=−12 the series in (9) takes the form
it is not a series in integral powers of x.
To investigate the possible existence of Frobenius series solutions, we begin with the equation
obtained by multiplying the equation in (3) by x2. If x=0 is a regular singular point, then p(x) and q(x) are analytic at x=0, so
Suppose that Eq. (10) has the Frobenius series solution
We may (and always do) assume that c0≠0 because the series must have a first nonzero term. Termwise differentiation in Eq. (12) leads to
and
Substitution of the series in Eqs. (11) through (14) in Eq. (10) now yields
Upon multiplying initial terms of the two products on the left-hand side here and then collecting coefficients of xr, we see that the lowest-degree term in Eq. (15) is c0[r(r−1)+p0r+q0]xr. If Eq. (15) is to be satisfied identically, then the coefficient of this term (as well as those of the higher-degree terms) must vanish. But we are assuming that c0≠0, so it follows that r must satisfy the quadratic equation
of precisely the same form as that obtained with the equidimensional equation in (7). Equation (16) is called the indicial equation of the differential equation in (10), and its two roots (possibly equal) are the exponents of the differential equation (at the regular singular point x=0).
Our derivation of Eq. (16) shows that if the Frobenius series y=xr∑cnxn is to be a solution of the differential equation in (10), then the exponent r must be one of the roots r1 and r2 of the indicial equation in (16). If r1≠r2, it follows that there are two possible Frobenius series solutions, whereas if r1=r2 there is only one possible Frobenius series solution; the second solution cannot be a Frobenius series. The exponents r1 and r2 in the possible Frobenius series solutions are determined (using the indicial equation) by the values p0=p(0) and q0=q(0) that we have discussed. In practice, particularly when the coefficients in the differential equation in the original form in (1) are polynomials, the simplest way of finding p0 and q0 is often to write the equation in the form
Then inspection of the series that appear in the two numerators reveals the constants p0 and q0.
Find the exponents in the possible Frobenius series solutions of the equation
We divide each term by 2x2(1+x) to recast the differential equation in the form
and thus see that p0=32 and q0=−12. Hence the indicial equation is
with roots r1=12 and r2=−1. The two possible Frobenius series solutions are then of the forms
Once the exponents r1 and r2 are known, the coefficients in a Frobenius series solution are determined by substitution of the series in Eqs. (12) through (14) in the differential equation, essentially the same method as was used to determine coefficients in power series solutions in Section 11.2. If the exponents r1 and r2 are complex conjugates, then there always exist two linearly independent Frobenius series solutions. We will restrict our attention here to the case in which r1 and r2 are both real. We also will seek solutions only for x>0. Once such a solution has been found, we need only replace xr1 with |x|r1 to obtain a solution for x<0. The following theorem is proved in Chapter 4 of Coddington’s An Introduction to Ordinary Differential Equations.
We have already seen that if r1=r2, then there can exist only one Frobenius series solution. It turns out that, if r1−r2 is a positive integer, then there may or may not exist a second Frobenius series solution of the form in Eq. (19) corresponding to the smaller root r2. Examples 4 through 6 illustrate the process of determining the coefficients in those Frobenius series solutions that are guaranteed by Theorem 1.
Find the Frobenius series solutions of
First we divide each term by 2x2 to put the equation in the form in (17):
We now see that x=0 is a regular singular point, and that p0=32 and q0=−12. Because p(x)≡32 and q(x)=−12−12x2 are polynomials, the Frobenius series we obtain will converge for all x>0. The indicial equation is
so the exponents are r1=12 and r2=−1. They do not differ by an integer, so Theorem 1 guarantees the existence of two linearly independent Frobenius series solutions. Rather than separately substituting
in Eq. (20), it is more efficient to begin by substituting y=xr∑cnxn. We will then get a recurrence relation that depends on r. With the value r1=12 it becomes a recurrence relation for the series for y1, whereas with r2=−1 it becomes a recurrence relation for the series for y2.
When we substitute
and
in Eq. (20)—the original differential equation, rather than Eq. (21)—we get
At this stage there are several ways to proceed. A good standard practice is to shift indices so that each exponent will be the same as the smallest one present. In this example, we shift the index of summation in the third sum by −2 to reduce its exponent from n+r+2 to n+r. This gives
The common range of summation is n≧2, so we must treat n=0 and n=1 separately. Following our standard practice, the terms corresponding to n=0 will always give the indicial equation
The terms corresponding to n=1 yield
Because the coefficient 2r2+5r+2 of c1 is nonzero whether r=12 or r=−1, it follows that
in either case.
The coefficient of xn+r in Eq. (23) is
We solve for cn and simplify to obtain the recurrence relation
Case 1: r1=12. We now write an in place of cn and substitute r=12 in Eq. (25). This gives the recurrence relation
With this formula we can determine the coefficients in the first Frobenius solution y1. In view of Eq. (24) we see that an=0 whenever n is odd. With n=2, 4, and 6 in Eq. (26), we get
Hence the first Frobenius solution is
Case 2: r2=−1. We now write bn in place of cn and substitute r=−1 in Eq. (25). This gives the recurrence relation
Again, Eq. (24) implies that bn=0 for n odd. With n=2, 4, and 6 in (27), we get
Hence the second Frobenius solution is
Find a Frobenius solution of Bessel’s equation of order zero,
In the form of (17), Eq. (28) becomes
Hence x=0 is a regular singular point with p(x)≡1 and q(x)=x2, so our series will converge for all x>0. Because p0=1 and q0=0, the indicial equation is
Thus we obtain only the single exponent r=0, and so there is only one Frobenius series solution
of Eq. (28); it is in fact a power series.
Thus we substitute y=∑cnxn in (28); the result is
We combine the first two sums and shift the index of summation in the third by −2 to obtain
The term corresponding to x0 gives 0=0: no information. The term corresponding to x1 gives c1=0, and the term for xn yields the recurrence relation
Because c1=0, we see that cn=0 whenever n is odd. Substituting n=2, 4, and 6 in Eq. (29), we get
Evidently, the pattern is
The choice c0=1 gives us one of the most important special functions in mathematics, the Bessel function of order zero of the first kind, denoted by J0(x). Thus
In this example we have not been able to find a second linearly independent solution of Bessel’s equation of order zero. We will derive that solution in Section 11.4; it will not be a Frobenius series.
Recall that, if r1−r2 is a positive integer, then Theorem 1 guarantees only the existence of the Frobenius series solution corresponding to the larger exponent r1. Example 6 illustrates the fortunate case in which the series method nevertheless yields a second Frobenius series solution. An example in which the second solution is not a Frobenius series will be discussed in Section 11.4.
Find the Frobenius series solutions of
In standard form the equation becomes
so we see that x=0 is a regular singular point with p0=2 and q0=0. The indicial equation
has roots r1=0 and r2=−1, which differ by an integer. In this case when r1−r2 is an integer, it is better to depart from the standard procedure of Example 4 and begin our work with the smaller exponent. As you will see, the recurrence relation will then tell us whether or not a second Frobenius series solution exists. If it does exist, our computations will simultaneously yield both Frobenius series solutions. If the second solution does not exist, we begin anew with the larger exponent r=r1 to obtain the one Frobenius series solution guaranteed by Theorem 1.
Hence we begin by substituting
in Eq. (31). This gives
We combine the first two sums and shift the index by −2 in the third to obtain
The cases n=0 and n=1 reduce to
Hence we have two arbitrary constants c0 and c1 and therefore can expect to find a general solution incorporating two linearly independent Frobenius series solutions. If, for n=1, we had obtained an equation such as 0·c1=3, which can be satisfied for no choice of c1, this would have told us that no second Frobenius series solution could exist.
Now knowing that all is well, from (32) we read the recurrence relation
The first few values of n give
evidently the pattern is
for n≧1. Therefore, a general solution of Eq. (31) is
Thus
We have thus found a general solution expressed as a linear combination of the two Frobenius series solutions
As indicated in Fig. 11.3.1, one of these Frobenius series solutions is bounded but the other is unbounded near the regular singular point x=0—a common occurrence in the case of exponents differing by an integer.
When confronted with a linear second-order differential equation
with analytic coefficient functions, in order to investigate the possible existence of series solutions we first write the equation in the standard form
If P(x) and Q(x) are both analytic at x=0, then x=0 is an ordinary point, and the equation has two linearly independent power series solutions.
Otherwise, x=0 is a singular point, and we next write the differential equation in the form
If p(x) and q(x) are both analytic at x=0, then x=0 is a regular singular point. In this case we find the two exponents r1 and r2 (assumed real, and with r1≧r2) by solving the indicial equation
where p0=p(0) and q0=q(0). There always exists a Frobenius series solution y=xr1∑anxn associated with the larger exponent r1, and if r1−r2 is not an integer, the existence of a second Frobenius series solution y2=xr2∑bnxn is also guaranteed.
In Problems 1 through 8, determine whether x=0 is an ordinary point, a regular singular point, or an irregular singular point. If it is a regular singular point, find the exponents of the differential equation at x=0.
xy″+(x−x3)y′+(sin x)y=0
xy″+x2y′+(ex−1)y=0
x2y″+(cos x)y′+xy=0
3x3y″+2x2y′+(1−x2)y=0
x(1+x)y″+2y′+3xy=0
x2(1−x2)y″+2xy′−2y=0
x2y″+(6sin x)y′+6y=0
(6x2+2x3)y″+21xy′+9(x2−1)y=0
If x=a≠0 is a singular point of a second-order linear differential equation, then the substitution t=x−a transforms it into a differential equation having t=0 as a singular point. We then attribute to the original equation at x=a the behavior of the new equation at t=0. Classify (as regular or irregular) the singular points of the differential equations in Problems 9 through 16.
(1−x)y″+xy′+x2y=0
(1−x)2y″+(2x−2)y′+y=0
(1−x2)y″−2xy′+12y=0
(x−2)3y″+3(x−2)2y′+x3y=0
(x2−4)y″+(x−2)y′+(x+2)y=0
(x2−9)2y″+(x2+9)y′+(x2+4)y=0
(x−2)2y″−(x2−4)y′+(x+2)y=0
x3(1−x)y″+(3x+2)y′+xy=0
Find two linearly independent Frobenius series solutions (for x>0) of each of the differential equations in Problems 17 through 26.
4xy″+2y′+y=0
2xy″+3y′−y=0
2xy″−y′−y=0
3xy″+2y′+2y=0
2x2y″+xy′−(1+2x2)y=0
2x2y″+xy′−(3−2x2)y=0
6x2y″+7xy′−(x2+2)y=0
3x2y″+2xy′+x2y=0
2xy″+(1+x)y′+y=0
2xy″+(1−2x2)y′−4xy=0
Use the method of Example 6 to find two linearly independent Frobenius series solutions of the differential equations in Problems 27 through 31. Then construct a graph showing their graphs for x>0.
xy″+2y′+9xy=0
xy″+2y′−4xy=0
4xy″+8y′+xy=0
xy″−y′+4x3y=0
4x2y″−4xy′+(3−4x2)y=0
In Problems 32 through 34, find the first three nonzero terms of each of two linearly independent Frobenius series solutions.
2x2y″+x(x+1)y′−(2x+1)y=0
(2x2+5x3)y″+(3x−x2)y′−(1+x)y=0
2x2y″+(sin x)y′−(cos x)y=0
Note that x=0 is an irregular point of the equation
Show that y=xr∑cnxn can satisfy this equation only if r=0.
Substitute y=∑cnxn to derive the “formal” solution y=∑n!xn. What is the radius of convergence of this series?
Suppose that A and B are nonzero constants. Show that the equation x2y″+Ay′+By=0 has at most one solution of the form y=xr∑cnxn.
Repeat part (a) with the equation x3y″+Axy′+By=0.
Show that the equation x3y″+Ax2y′+By=0 has no Frobenius series solution. (Suggestion: In each case substitute y=xr∑cnxn in the given equation to determine the possible values of r.)
Use the method of Frobenius to derive the solution y1=x of the equation x3y″−xy′+y=0.
Verify by substitution the second solution y2=xe−1/x. Does y2 have a Frobenius series representation?
Apply the method of Frobenius to Bessel’s equation of order 12,
to derive its general solution for x>0,
Figure 11.3.2 shows the graphs of the two indicated solutions.
Show that Bessel’s equation of order 1,
has exponents r1=1 and r2=−1 at x=0, and that the Frobenius series corresponding to r1=1 is
Show that there is no Frobenius solution corresponding to the smaller exponent r2=−1;that is, show that it is impossible to determine the coefficients in
Consider the equation x2y″+xy′+(1−x)y=0.
Show that its exponents are ±i, so it has complex-valued Frobenius series solutions
with p0=q0=1.
Show that the recursion formula is
Apply this formula with r=i to obtain pn=cn, then with r=−i to obtain qn=cn. Conclude that pn and qn are complex conjugates: pn=an+ibn and qn=an−ibn, where the numbers {an} and {bn} are real.
Deduce from part (b) that the differential equation given in this problem has real-valued solutions of the form
where A(x)=∑anxn and B(x)=∑bnxn.
Consider the differential equation
that appeared in an advertisement for a symbolic algebra program in the March 1984 issue of the American Mathematical Monthly.
Show that x=0 is a regular singular point with exponents r1=1 and r2=0.
It follows from Theorem 1 that this differential equation has a power series solution of the form
Substitute this series (with c1=1) in the differential equation to show that c2=−2, c3=3, and
for n≧2.
Use the recurrence relation in part (b) to prove by induction that cn=(−1)n+1n for n≧1 (!). Hence deduce (using the geometric series) that
for 0<x<1.
This problem is a brief introduction to Gauss’s hypergeometric equation
where α, β, and γ are constants. This famous equation has wide-ranging applications in mathematics and physics.
Show that x=0 is a regular singular point of Eq. (35), with exponents 0 and 1−γ.
If γ is not zero or a negative integer, it follows (why?) that Eq. (35) has a power series solution
with c0≠0. Show that the recurrence relation for this series is
for n≧0.
Conclude that with c0=1 the series in part (b) is
where αn=α(α+1)(α+2)⋯(α+n−1) for n≧1, and βn and γn are defined similarly.
The series in (36) is known as the hypergeometric series and is commonly denoted by F(α,β,γ,x). Show that
F(1,1,1,x)=11−x (the geometric series);
xF(1,1,2,−x)=ln (1+x);
xF(12,1,32,−x2)=tan−1x;
F(−k,1,1,−x)=(1+x)k (the binomial series).
Here we illustrate the use of a computer algebra system such as Maple to apply the method of Frobenius. More complete versions of this application—illustrating the use of Maple, Mathematica, and Matlab—can be downloaded from our Expanded Applications site using the URL indicated in the margin. We consider the differential equation
of Example 4 in this section, where we found the two indicial roots r1=12 and r2=−1.
Beginning with the indicial root r1=12, we first write the initial seven terms of a proposed Frobenius series solution:
a := array(0..6):
y := x^(1/2) ∗sum( a[n] ∗x^(n), n = 0..6);
Then we substitute this series (actually, partial sum) into the left-hand side of Eq. (1).
deq1 := 2∗x^2∗diff(y,x$2) + 3∗x∗diff(y,x) − (x^2 + 1)∗y:
Evidently x3/2 will factor out upon simplification, so we multiply by x−3/2 and then collect coefficients of like powers of x.
deq2 := collect( x^(−3/2)∗simplify(deq1), x);
We see here the equations that the successive coefficients must satisfy. We can select them automatically by defining an array, then filling the elements of this array by equating to zero (in turn)each of the coefficients in the series.
eqs := array(0..5):
for n from 0 to 5 do
eqs[n] := coeff(deq1,x,n) = 0: od:
coeffEqs := convert(eqs, set);
We now have a collection of six linear equations relating the seven coefficients (a0 through a6). Hence we can proceed to solve for the successive coefficients in terms of a0.
succCoeffs := convert([seq(a[n], n=1..6)], set);
ourCoeffs := solve(coeffEqs, succCoeffs);
Thus we get the first particular solution
found in Example 4. You can now repeat this process, beginning with the indicial root r2=−1, to derive similarly the second particular solution.
In the following problems, use this method to derive Frobenius series solutions that can be checked against the given known general solutions.
xy″−y′+4x3y=0,y(x)=Acosx2+Bsinx2
xy″−2y′+9x5y=0,y(x)=Acosx3+Bsinx3
4xy″−2y′+y=0,y(x)=Acos√x+Bsin√x
xy″+2y′+xy=0,y(x)=1x(Acos x+Bsin x)
4xy″+6y′+y=0,y(x)=1√x(Acos√x+Bsin√x)
x2y″+xy′+(4x4−1)y=0,y(x)=1x(Acosx2+Bsinx2)
xy″+3y′+4x3y=0,y(x)=1x2(Acosx2+Bsinx2)
x2y″+x2y′−2y=0,y(x)=1x[A(2−x)+B(2+x)e−x]
Problems 9 through 11 involve the arctangent series
(x+x3)y″+(2+4x2)y′+2xy=0,y(x)=1x(A+Btan−1x)
(2x+2x2)y″+(3+5x)y′+y=0,y(x)=1√x(A+Btan−1√x)
(x+x5)y″+(3+7x4)y′+8x3y=0,y(x)=1x2(A+Btan−1x2)
18.188.38.142