ﬁrst applies it to the problem of determining the lowestfrequency of vibration of a stretched string. His argumentand conclusions are presented here in a modernizednotation.Consider a string length
l
with a constant mass density
r
and tension
t
. The string is supposed to exist along theline segment
À
l
/2
r
x
r
l
/2, and its height at position
x
andtime
t
will be denoted by
y
(
x,t
). Now, a natural mode of thestring occurs when
y
(
x,t
) is nonzero in the absence of anexternal driving force. Since the natural modes of losslesslinear systems are always sinusoidal, the displacement of the string may be taken to be
y
ð
x
;
t
Þ¼
v
ð
x
Þ
cos
o
t
ð
1
Þ
The total kinetic energy of the string can be expressed as
T
ð
t
Þ¼
12
Z
l
=
2
Àð
l
=
2
Þ
r
@
y
@
t
2
dx
¼
o
2
sin
2
o
t
½
~
T T
ð
v
Þ ð
2
Þ
where
~
T T
ð
f
Þ¼
12
Z
l
=
2
Àð
l
=
2
Þ
r
½
v
ð
x
Þ
2
dx
ð
3
Þ
Similarly, the total potential energy of the string may bewritten as
V
ð
t
Þ¼
12
Z
l
=
2
Àð
l
=
2
Þ
t
@
y
@
x
2
dx
¼
cos
2
o
t
½
~
V V
ð
v
Þ ð
4
Þ
where
~
V V
ð
f
Þ¼
12
Z
l
=
2
Àð
l
=
2
Þ
t
½
v
0
ð
x
Þ
2
dx
ð
5
Þ
and the prime denotes differentiation.The requirement that the energy of the system remainconstant implies that
ddt
½
T
ð
t
Þþ
V
ð
t
Þ¼
0
ð
6
Þ
So long as
o
a
0 (which indicates a trivial solution any-way), the value of
o
corresponding to a given
v
(
x
) thatsolves the appropriate wave equation is given by
o
2
¼
~
V V
ð
v
Þ
~
T T
ð
v
Þð
7
Þ
In the paragraph preceding his derivation of this result(which was slightly different than that presented here),Rayleigh made a heuristic argument for discrete problemsthat demonstrates that the minimum positive value of
o
delivered by this expression (over all admissible functions
v
) is in fact the minimum frequency offree vibration of thesystem. He then suggested that formulas of the form of Eq. (7) might be useful for computing the minimum freevibration frequency of systems. Specifically, if a function
v
(
x
) is a good approximation to the true solution, in errorby a linear quantity
D
v
, then the approximation of
o
2
willbe in error only by a quantity of order (
D
v
)
2
. Moreover, if aset of functions
~
uu
n
ð
x
Þ
is substituted into Eq. (7), the bestestimate of
o
2
is necessarily the smallest.For the problem given above, the true solution is foundwhen
v
ð
x
Þ¼
u
ð
x
Þ¼
cos
p
xl
ð
8
Þ
for which
o
2
¼
p
2
tr
l
2
ð
9
Þ
(We will always use the variable
v
to indicate arbitraryfunctions, and the variable
u
to indicate true solutions.Trial solutions are indicated by
~
uu
n
, where the subscriptindicates some parameter deﬁning the trial.) Rayleighsuggested testing the method with the set of functions
~
uu
n
ð
x
Þ¼
1
À
2
j
x
j
l
n
ð
10
Þ
for
n
Z
1, for which
~
T T
ð
~
uu
n
Þ¼
r
ln
2
ð
n
þ
1
Þð
2
n
þ
1
Þð
11
Þ
and
~
V V
ð
~
uu
n
Þ¼
2
t
n
2
ð
2
n
À
1
Þ
l
ð
12
Þ
These expressions lead to an approximation of the form
o
2
n
¼
2
ð
n
þ
1
Þð
2
n
þ
1
Þ
2
n
À
1
tr
l
2
¼
l
2
n
tr
l
2
ð
13
Þ
From this expression,
l
1
¼
ﬃﬃﬃﬃﬃﬃ
12
p
, off from the true value of
p
by a little more than 10%. Similarly,
l
2
¼
ﬃﬃﬃﬃﬃﬃ
10
p
, a valuein error by 0.65%. The best possible value of
n
is the onethat minimizes Eq. (13); this value is
n
opt
¼
12
ð
1
þ
ﬃﬃﬃ
6
p Þ%
1
:
72474. Using this value in Eq. (13)gives rise to the value
l
n
opt
¼
3
:
1463 and a scant error of 0.149%.Throughout the rest of
The Theory of Sound
, Rayleighcontinues to apply this method to problems that cannot besolved analytically, especially problems involving inhomo-geneous material parameters. While Rayleigh’s method aspresented above has been superceded by the Ritz (i.e.,Rayleigh–Ritz) method described below, Rayleigh quo-tients similar to Eq. (7) are still used today in numericallinear algebra in the eigenvalue computation techniqueknown as
Rayleigh quotient iteration
[14].Rayleigh’s method was generalized by the Swiss phys-icist Walther Ritz shortly before he died of tuberculosis in1909 [15]. Specifically, Ritz proposed a method for thesolution of physical problems expressed variationally;
1736 GALERKIN METHOD (RAYLEIGH–RITZ METHOD)