SturmLiouville theory

The eigenvalues 1 , 2 , 3 , ... of the regular Sturm

Liouville problem (1)-(2)-(3) are real and can be ordered such that

In mathematics and its applications, a classical Sturm

Liouville equation, named after Jacques Charles
Franois Sturm (18031855) and Joseph Liouville
(18091882), is a real second-order linear dierential
equation of the form

1 < 2 < 3 < < n < ;

Corresponding to each eigenvalue n is a unique
(up to a normalization constant) eigenfunction yn(x)
which has exactly n 1 zeros in (a, b). The eigenfunction yn(x) is called the n-th fundamental solution satisfying the regular SturmLiouville problem
(1)-(2)-(3).

where y is a function of the free variable x. Here the functions p(x), q(x), and w(x) > 0 are specied at the outset. In
the simplest of cases all coecients are continuous on the
nite closed interval [a,b], and p has continuous derivative. In this simplest of all cases, this function y is called
a solution if it is continuously dierentiable on (a,b) and
satises the equation (1) at every point in (a,b). In addition, the unknown function y is typically required to
satisfy some boundary conditions at a and b. The function w(x), which is sometimes called r(x), is called the
weight or density function.

The normalized eigenfunctions form an orthonormal

basis

yn (x)ym (x)w(x) dx = mn ,

The value of is not specied in the equation; nding the

values of for which there exists a non-trivial solution
of (1) satisfying the boundary conditions is part of the
problem called the SturmLiouville (SL) problem.

in the Hilbert space L2 ([a, b], w(x)dx). Here

mn is a Kronecker delta.

Such values of , when they exist, are called the

eigenvalues of the boundary value problem dened by (1)
and the prescribed set of boundary conditions. The corresponding solutions (for such a ) are the eigenfunctions
of this problem. Under normal assumptions on the coefcient functions p(x), q(x), and w(x) above, they induce
a Hermitian dierential operator in some function space
dened by boundary conditions. The resulting theory of
the existence and asymptotic behavior of the eigenvalues,
the corresponding qualitative theory of the eigenfunctions
and their completeness in a suitable function space became known as SturmLiouville theory. This theory is
important in applied mathematics, where SL problems
occur very commonly, particularly when dealing with linear partial dierential equations that are separable.

Note that, unless p(x) is continuously dierentiable and

q(x), w(x) are continuous, the equation has to be understood in a weak sense.

1 SturmLiouville form
The dierential equation (1) is said to be in Sturm
Liouville form or self-adjoint form. All second-order
linear ordinary dierential equations can be recast in the
form on the left-hand side of (1) by multiplying both sides
of the equation by an appropriate integrating factor (although the same is not true of second-order partial differential equations, or if y is a vector.)

A SturmLiouville (SL) problem is said to be regular if

p(x), w(x) > 0, and p(x), p'(x), q(x), and w(x) are con- 1.1 Examples
tinuous functions over the nite interval [a, b], and have
separated boundary conditions of the form
1.1.1 The Bessel equation
(
)
x2 y + xy + x2 2 y = 0
which can be written in SturmLiouville form as
(
)
Under the assumption that the SL problem is regular,
2

(xy ) + x
y = 0.
the main tenet of SturmLiouville theory states that:
x
1

1.1.2

STURMLIOUVILLE EQUATIONS AS SELF-ADJOINT DIFFERENTIAL OPERATORS

The Legendre equation

(1 x2 )y 2xy + ( + 1)y = 0

2 SturmLiouville equations as
self-adjoint dierential operators

which can easily be put into SturmLiouville form, since

D(1 x2 ) = 2x, so, the Legendre equation is equivalent
The map
to
[(1 x2 )y ] + ( + 1)y = 0
1.1.3

Lu =

An example using an integrating factor

x3 y xy + 2y = 0.
Divide throughout by x3 :
1
2
y + 3y = 0
x2
x
Multiplying throughout by an integrating factor of
y

(x) = e

x12 dx

= ex ,

gives
1

ex
x2
so the dierential equation is equivalent to
1

De x =

1.1.4

2e x
y = 0.
x3

The integrating factor for a general second order dierential equation

P (x)y + Q(x)y + R(x)y = 0

multiplying through by the integrating factor

(x) =

1
e
P (x)

Q(x)
dx
P (x)

and then collecting gives the SturmLiouville form:

d
((x)P (x)y ) + (x)R(x)y = 0
dx
or, explicitly,
d (
e
dx

Q(x)
dx
P (x)

) R(x)
e
y +
P (x)

Q(x)
P (x)

dx

y=0

[
]
)
d
du
p(x)
+ q(x)u
dx
dx

can be viewed as a linear operator mapping a function u to

another function Lu. One may study this linear operator
in the context of functional analysis. In fact, equation (1)
can be written as

Lu = u.
This is precisely the eigenvalue problem; that is, one is
trying to nd the eigenvalues 1 , 2 , 3 , ... and the corresponding eigenvectors u1 , u2 , u3 , ... of the L operator.
The proper setting for this problem is the Hilbert space
L2 ([a, b], w(x) dx) with scalar product

ex
2e x
e y 2 y + 3 y = 0
x
x
which can be easily put into SturmLiouville form since
1
x

(e x y ) +

1
w(x)

f, g =

f (x)g(x)w(x) dx.
a

In this space L is dened on suciently smooth functions

which satisfy the above boundary conditions. Moreover,
L gives rise to a self-adjoint operator. This can be seen
formally by using integration by parts twice, where the
boundary terms vanish by virtue of the boundary conditions. It then follows that the eigenvalues of a Sturm
Liouville operator are real and that eigenfunctions of L
corresponding to dierent eigenvalues are orthogonal.
However, this operator is unbounded and hence existence
of an orthonormal basis of eigenfunctions is not evident.
To overcome this problem, one looks at the resolvent

(L z)1 ,

z C,

where z is chosen to be some real number which is not

an eigenvalue. Then, computing the resolvent amounts to
solving the inhomogeneous equation, which can be done
using the variation of parameters formula. This shows
that the resolvent is an integral operator with a continuous symmetric kernel (the Greens function of the problem). As a consequence of the ArzelAscoli theorem,
this integral operator is compact and existence of a sequence of eigenvalues n which converge to 0 and eigenfunctions which form an orthonormal basis follows from
the spectral theorem for compact operators. Finally, note
that

(L z)1 u = u,

(
)
Lu = z + 1 u,

3
are equivalent.

which is all we need for this particular theory to function. We mention for the interested reader that in this
case we may rely on a result which says that Fouriers
series converges at every point of dierentiability, and
at jump points (the function x, considered as a periodic
function, has a jump at ) converges to the average of the
left and right limits (see convergence of Fourier series).

If the interval is unbounded, or if the coecients have

singularities at the boundary points, one calls L singular.
In this case, the spectrum no longer consists of eigenvalues alone and can contain a continuous component. There
is still an associated eigenfunction expansion (similar to
Fourier series versus Fourier transform). This is important in quantum mechanics, since the one-dimensional Therefore, by using formula (4), we obtain that the solutime-independent Schrdinger equation is a special case tion is
of a SL equation.

Example

u=

(1)k
2 3 sin kx.
k

k=1

We wish to nd a function u(x) which solves the following In this case, we could have found the answer using antidierentiation. This technique yields
SturmLiouville problem:

u=

1
6

( 3
)
x 2 x ,

where the unknowns are and u(x). As above, we must

whose Fourier series agrees with the solution we found.
add boundary conditions, we take for example
The anti-dierentiation technique is no longer useful in
most cases when the dierential equation is in many variables.
u(0) = u() = 0.
Observe that if k is any integer, then the function

u(x) = sin kx
is a solution with eigenvalue = k2 . We know that the
solutions of a SL problem form an orthogonal basis, and
we know from Fourier series that this set of sinusoidal
functions is an orthogonal basis. Since orthogonal bases
are always maximal (by denition) we conclude that the
SL problem in this case has no other eigenvectors.
Given the preceding, let us now solve the inhomogeneous
problem

4 Application to normal modes

Certain partial dierential equations can be solved with
the help of SL theory. Suppose we are interested in the
modes of vibration of a thin membrane, held in a rectangular frame, 0 x L1 , 0 y L2 . The equation of
motion for the vertical membranes displacement, W(x,
y, t) is given by the wave equation:

2W
1 2W
2W
+
=
.
x2
y 2
c2 t2

The method of separation of variables suggests looking

rst for solutions of the simple form W = X(x) Y(y)
Lu = x,
x (0, )
T(t). For such a function W the partial dierential equation becomes X"/X + Y/Y = (1/c2 )T/T. Since the three
with the same boundary conditions. In this case, we must terms of this equation are functions of x,y,t separately,
write f(x) = x in a Fourier series. The reader may check, they must be constants. For example, the rst term gives
either by integrating exp(ikx)x dx or by consulting a table X" = X for a constant . The boundary conditions (held
of Fourier transforms, that we thus obtain
in a rectangular frame) are W = 0 when x = 0, L1 or y
= 0, L2 and dene the simplest possible SL eigenvalue
problems
as in the example, yielding the normal mode

(1)k
solutions
for W with harmonic time dependence,
sin kx.
Lu =
2
k
k=1

) (
)
(
This particular Fourier series is troublesome because of
ny
mx
sin
cos (mn t)
its poor convergence properties. It is not clear a pri- Wmn (x, y, t) = Amn sin L1
L2
ori whether the series converges pointwise. Because
of Fourier analysis, since the Fourier coecients are where m and n are non-zero integers, Amn are arbitrary
"square-summable", the Fourier series converges in L2 constants, and

6 APPLICATION TO PDES

(
2
mn
= c2

2 2

2 2

n
m
+
L21
L22

)
.

when n > 0. The resulting iterated integrals are now applied as coecients in the following two power series in
:

The functions Wmn form a basis for the Hilbert space of

(generalized) solutions of the wave equation; that is, an

k e (2k)
,
u
=
y
( 0 ) X
0
0
arbitrary solution W can be decomposed into a sum of
k=0
these modes, which vibrate at their individual frequencies mn . This representation may require a convergent

k
innite sum.
u1 = y0
( 0 ) X (2k+1) .
k=0

Representation of solutions and

numerical calculation

Then for any (real or complex), u0 and u1 are linearly

independent solutions of the corresponding equation (1).
(The functions p(x) and q(x) take part in this construction
through their inuence on the choice of y0 .)

Next one chooses coecients c0 , c1 so that the combination y = c0 u0 + c1 u1 satises the rst boundary condition
(2). This is simple to do since X(n) (a) = 0 and X~(n) (a) =
0, for n > 0. The values of X(n) (b) and X~(n) (b) provide
the values of u0 (b) and u1 (b) and the derivatives u0 '(b)
and u1 '(b), so the second boundary condition (3) becomes
an equation in a power series in . For numerical work
1. Shooting methods.[1][2] These methods proceed by
one may truncate this series to a nite number of terms,
guessing a value of , solving an initial value problem
producing a calculable polynomial in whose roots are
dened by the boundary conditions at one endpoint, say,
approximations of the sought-after eigenvalues.
a, of the interval [a, b], comparing the value this solution takes at the other endpoint b with the other desired When = 0 , this reduces to the original construction
boundary condition, and nally increasing or decreasing described above for a solution linearly independent to a
as necessary to correct the original value. This strategy given one. The representations (5),(6) also have theoretical applications in SturmLiouville theory.[3]
is not applicable for locating complex eigenvalues.
The SturmLiouville dierential equation (1) with
boundary conditions may be solved in practice by a variety of numerical methods. In dicult cases, one may
need to carry out the intermediate calculations to several
hundred decimal places of accuracy in order to obtain the
eigenvalues correctly to a few decimal places.

2. Finite dierence method.

3.
The Spectral Parameter Power Series (SPPS) 5.1 Construction of a nonvanishing solumethod[3] makes use of a generalization of the following
tion
fact about second order ordinary dierential equations:
if y is a solution which does not vanish at any point of The SPPS method can, itself, be used to nd a starting
[a,b], then the function
solution y0 . Consider the equation (py')' = qy; i.e., q,
w, and are replaced in (1) by 0, q, and respectively.
Then
the constant function 1 is a nonvanishing solution
x
dt
corresponding
to the eigenvalue 0 = 0. While there is no
y(x)
2
a p(t)y(t)
guarantee that u0 or u1 will not vanish, the complex funcis a solution of the same equation and is linearly indepen- tion y0 = u0 + iu1 will never vanish because two linearly
dent from y. Further, all solutions are linear combinations independent solutions of a regular SL equation cannot
of these two solutions. In the SPPS algorithm, one must vanish simultaneously as a consequence of the Sturm sepbegin with an arbitrary value 0 * (often 0 * = 0; it does aration theorem. This trick gives a solution y0 of (1) for
not need to be an eigenvalue) and any solution y0 of (1) the value 0 = 0. In practice if (1) has real coecients,
with = 0 * which does not vanish on [a, b]. (Discus- the solutions based on y0 will have very small imaginary
sion below of ways to nd appropriate y0 and 0 * .) Two parts which must be discarded.

sequences of functions X(n) (t), X~(n) (t) on [a, b], referred

to as iterated integrals, are dened recursively as follows.
First when n = 0, they are taken to be identically equal 6 Application to PDEs
to 1 on [a, b]. To obtain the next functions they are multiplied alternately by 1/(py0 2 ) and wy0 2 and integrated, For a linear second order in one spatial dimension and
specically
rst order in time of the form:

f (x)

u
u
2u
+ g(x)
+ h(x)u =
+ k(t)u
x2
x
t

7 See also

u(a, t) = u(b, t) = 0
u(x, 0) = s(x)
Let us apply separation of variables, which in doing we
must impose that:

Normal mode
Oscillation theory
Self-adjoint

u(x, t) = X(x)T (t)

Variation of parameters

Then our above PDE may be written as:

Spectral theory of ordinary dierential equations

AtkinsonMingarelli theorem

T (t)

M
LX(x)
=
X(x)
T (t)

8 References

Where
[1] J. D. Pryce, Numerical Solution of SturmLiouville Problems, Clarendon Press, Oxford, 1993.
2
= f (x) d + g(x) d + h(x),
L
dx2
dx

= d + k(t)
M
dt

and X(x) are independent of time

Since, by denition, L
and T (t) are independent of position x, then both
t and M
sides of the above equation must be equal to a constant:

LX(x)
= X(x)

[2] V. Ledoux, M. Van Daele, G. Vanden Berghe, Ecient

computation of high index SturmLiouville eigenvalues
for problems in physics, Comput. Phys. Comm. 180,
2009, 532554.
[3] V. V. Kravchenko, R. M. Porter, Spectral parameter
power series for SturmLiouville problems, Mathematical Methods in the Applied Sciences (MMAS) 33, 2010,
459468

X(a) = X(b) = 0

9 Further reading

T (t) = T (t)
M
The rst of these equations must be solved as a Sturm
Liouville problem. Since there is no general analytic (exact) solution to SturmLiouville problems, we can assume
we already have the solution to this problem, that is, we
have the eigenfunctions Xn (x) and eigenvalues n . The
second of these equations can be analytically solved once
the eigenvalues are known.

d
Tn (t) = (n k(t))Tn (t)
dt
t

Tn (t) = an e(n t 0 k( )d )
t

an Xn (x)e(n t 0 k( )d )
u(x, t) =
n

an =

Xn (x), s(x)
Xn (x), Xn (x)

Where:

y(x), z(x) =

y(x)z(x)w(x)dx

Hazewinkel, Michiel, ed. (2001), Sturm-Liouville

theory, Encyclopedia of Mathematics, Springer,
ISBN 978-1-55608-010-4
Hartman, Philip (2002). Ordinary Dierential
Equations (2 ed.). Philadelphia: SIAM. ISBN 9780-89871-510-1.
Polyanin, A. D. and Zaitsev, V. F. (2003). Handbook of Exact Solutions for Ordinary Dierential Equations (2 ed.). Boca Raton: Chapman &
Hall/CRC Press. ISBN 1-58488-297-2.
Teschl, Gerald (2012).
Ordinary Dierential
Equations and Dynamical Systems. Providence:
American Mathematical Society. ISBN 978-08218-8328-0. (Chapter 5)
Teschl, Gerald (2009).
Mathematical Methods in Quantum Mechanics; With Applications to
Schrdinger Operators. Providence: American
Mathematical Society. ISBN 978-0-8218-4660-5.
(see Chapter 9 for singular SL operators and connections with quantum mechanics)

w(x) =

g(x)
dx
f (x)

f (x)

Zettl, Anton (2005). SturmLiouville Theory.

Providence: American Mathematical Society.
ISBN 0-8218-3905-5.

9 FURTHER READING
Birkho, Garrett (1973). A source book in classical
analysis. Cambridge, Massachusetts: Harvard University Press. ISBN 0-674-82245-5. (See Chapter
8, part B, for excerpts from the works of Sturm and
Liouville and commentary on them.)

10
10.1

Text and image sources, contributors, and licenses

Text

SturmLiouville theory Source: http://en.wikipedia.org/wiki/Sturm%E2%80%93Liouville_theory?oldid=656853057 Contributors:

Gareth Owen, Michael Hardy, Loisel, Stevenj, Charles Matthews, Dysprosia, Jitse Niesen, MathMartin, Wile E. Heresiarch, Tobias Bergemann, Pavon, Giftlite, MathKnight, FrozenUmbrella, Abdull, Esperant, Shotwell, Gadykozma, Xmath, Killing Vector, Oleg Alexandrov,
Grammarbot, Rjwilmsi, Hgkamath, Andrei Polyanin, Chubby Chicken, Sodin, WriterHound, YurikBot, RussBot, JabberWok, ArsniureDeGallium, Larsobrien, Rwalker, SmackBot, Edgar181, Xaosux, Amingare, Courcelles, WISo, Thijs!bot, Danger, Bencherlite, Mike409,
Paulnwatts, Pope Tony Liu, Gulloar, Gene Ray (Cubic), Pope Nigel the porter, Adam the brute, Policron, Quiet Silent Bob, Tony X Liu,
VolkovBot, Kyle the bot, TXiKiBoT, Dojarca, Zurishaddai, Uetz, Boatsstock, VVVBot, Richard Molnr-Szipai, Mr Travelling Wave Tube,
Emperor Tony X. Liu III, Emperor Tony X. Liu IV, Emperor Tony X. Liu VI, Emperor Tony X. Liu VII, Chokoboii, Warbler271, Mild Bill
Hiccup, Justin545, Addbot, Favonian, Wikomidia, Meisam, Yobot, Parkyere, J04n, RibotBOT, Sawomir Biay, Neutiquam, Klemensk,
Debator of mathematics, EmausBot, Tommy2010, Netheril96, ZroBot, RMPK, Mathuvw, Wcherowi, Helpful Pixie Bot, Wikispaghetti,
Meldraft, Randomguess, Fuse809, Makecat-bot, DennisMYeh, Smota123,
and Anonymous: 67

10.2

Images

10.3

Content license

Creative Commons Attribution-Share Alike 3.0