Chebyshev Polynomials

Reading Problems
Dierential Equation and Its Solution
The Chebyshev dierential equation is written as
(1 x
2
)
d
2
y
dx
2
x
dy
dx
+ n
2
y = 0 n = 0, 1, 2, 3, . . .
If we let x = cos t we obtain
d
2
y
dt
2
+ n
2
y = 0
whose general solution is
y = Acos nt + Bsin nt
or as
y = Acos(ncos
1
x) + Bsin(ncos
1
x) |x| < 1
or equivalently
y = AT
n
(x) + BU
n
(x) |x| < 1
where T
n
(x) and U
n
(x) are dened as Chebyshev polynomials of the rst and second kind
of degree n, respectively.
1
If we let x = cosh t we obtain
d
2
y
dt
2
n
2
y = 0
whose general solution is
y = Acosh nt + Bsinh nt
or as
y = Acosh(ncosh
1
x) + Bsinh(ncosh
1
x) |x| > 1
or equivalently
y = AT
n
(x) + BU
n
(x) |x| > 1
The function T
n
(x) is a polynomial. For |x| < 1 we have
T
n
(x) + iU
n
(x) = (cos t + i sin t)
n
=
_
x + i
_
1 x
2
_
n
T
n
(x) iU
n
(x) = (cos t i sin t)
n
=
_
x i
_
1 x
2
_
n
from which we obtain
T
n
(x) =
1
2
__
x + i
_
1 x
2
_
n
+
_
x i
_
1 x
2
_
n
_
For |x| > 1 we have
T
n
(x) + U
n
(x) = e
nt
=
_
x
_
x
2
1
_
n
T
n
(x) U
n
(x) = e
nt
=
_
x
_
x
2
1
_
n
The sum of the last two relationships give the same result for T
n
(x).
2
Chebyshev Polynomials of the First Kind of Degree n
The Chebyshev polynomials T
n
(x) can be obtained by means of Rodrigues formula
T
n
(x) =
(2)
n
n!
(2n)!
_
1 x
2
d
n
dx
n
(1 x
2
)
n1/2
n = 0, 1, 2, 3, . . .
The rst twelve Chebyshev polynomials are listed in Table 1 and then as powers of x in
terms of T
n
(x) in Table 2.
3
Table 1: Chebyshev Polynomials of the First Kind
T
0
(x) = 1
T
1
(x) = x
T
2
(x) = 2x
2
1
T
3
(x) = 4x
3
3x
T
4
(x) = 8x
4
8x
2
+ 1
T
5
(x) = 16x
5
20x
3
+ 5x
T
6
(x) = 32x
6
48x
4
+ 18x
2
1
T
7
(x) = 64x
7
112x
5
+ 56x
3
7x
T
8
(x) = 128x
8
256x
6
+ 160x
4
32x
2
+ 1
T
9
(x) = 256x
9
576x
7
+ 432x
5
120x
3
+ 9x
T
10
(x) = 512x
10
1280x
8
+ 1120x
6
400x
4
+ 50x
2
1
T
11
(x) = 1024x
11
2816x
9
+ 2816x
7
1232x
5
+ 220x
3
11x
4
Table 2: Powers of x as functions of T
n
(x)
1 = T
0
x = T
1
x
2
=
1
2
(T
0
+ T
2
)
x
3
=
1
4
(3T
1
+ T
3
)
x
4
=
1
8
(3T
0
+ 4T
2
+ T
4
)
x
5
=
1
16
(10T
1
+ 5T
3
+ T
5
)
x
6
=
1
32
(10T
0
+ 15T
2
+ 6T
4
+ T
6
)
x
7
=
1
64
(35T
1
+ 21T
3
+ 7T
5
+ T
7
)
x
8
=
1
128
(35T
0
+ 56T
2
+ 28T
4
+ 8T
6
+ T
8
)
x
9
=
1
256
(126T
1
+ 84T
3
+ 36T
5
+ 9T
7
+ T
9
)
x
10
=
1
512
(126T
0
+ 210T
2
+ 120T
4
+ 45T
6
+ 10T
8
+ T
10
)
x
11
=
1
1024
(462T
1
+ 330T
3
+ 165T
5
+ 55T
7
+ 11T
9
+ T
11
)
5
Generating Function for T
n
(x)
The Chebyshev polynomials of the rst kind can be developed by means of the generating
function
1 tx
1 2tx + t
2
=

n=0
T
n
(x)t
n
Recurrence Formulas for T
n
(x)
When the rst two Chebyshev polynomials T
0
(x) and T
1
(x) are known, all other polyno-
mials T
n
(x), n 2 can be obtained by means of the recurrence formula
T
n+1
(x) = 2xT
n
(x) T
n1
(x)
The derivative of T
n
(x) with respect to x can be obtained from
(1 x
2
)T

n
(x) = nxT
n
(x) + nT
n1
(x)
Special Values of T
n
(x)
The following special values and properties of T
n
(x) are often useful:
T
n
(x) = (1)
n
T
n
(x) T
2n
(0) = (1)
n
T
n
(1) = 1 T
2n+1
(0) = 0
T
n
(1) = (1)
n
6
Orthogonality Property of T
n
(x)
We can determine the orthogonality properties for the Chebyshev polynomials of the rst
kind from our knowledge of the orthogonality of the cosine functions, namely,
_

0
cos(m) cos(n ) d =
_

_
0 (m = n)
/2 (m = n = 0)
(m = n = 0)
Then substituting
T
n
(x) = cos(n)
cos = x
to obtain the orthogonality properties of the Chebyshev polynomials:
_
1
1
T
m
(x) T
n
(x) dx

1 x
2
=
_

_
0 (m = n)
/2 (m = n = 0)
(m = n = 0)
We observe that the Chebyshev polynomials form an orthogonal set on the interval 1
x 1 with the weighting function (1 x
2
)
1/2
Orthogonal Series of Chebyshev Polynomials
An arbitrary function f(x) which is continuous and single-valued, dened over the interval
1 x 1, can be expanded as a series of Chebyshev polynomials:
f(x) = A
0
T
0
(x) + A
1
T
1
(x) + A
2
T
2
(x) + . . .
=

n=0
A
n
T
n
(x)
7
where the coecients A
n
are given by
A
0
=
1

_
1
1
f(x) dx

1 x
2
n = 0
and
A
n
=
2

_
1
1
f(x) T
n
(x) dx

1 x
2
n = 1, 2, 3, . . .
The following denite integrals are often useful in the series expansion of f(x):
_
1
1
dx

1 x
2
=
_
1
1
x
3
dx

1 x
2
= 0
_
1
1
x dx

1 x
2
= 0
_
1
1
x
4
dx

1 x
2
=
3
8
_
1
1
x
2
dx

1 x
2
=

2
_
1
1
x
5
dx

1 x
2
= 0
Chebyshev Polynomials Over a Discrete Set of Points
A continuous function over a continuous interval is often replaced by a set of discrete values
of the function at discrete points. It can be shown that the Chebyshev polynomials T
n
(x)
are orthogonal over the following discrete set of N + 1 points x
i
, equally spaced on ,

i
= 0,

N
,
2
N
, . . . (N 1)

N
,
where
x
i
= arccos
i
We have
8
1
2
T
m
(1)T
n
(1)+
N1

i=2
T
m
(x
i
)T
n
(x
i
)+
1
2
T
m
(1)T
n
(1) =
_

_
0 (m = n)
N/2 (m = n = 0)
N (m = n = 0)
The T
m
(x) are also orthogonal over the following N points t
i
equally spaced,

i
=

2N
,
3
2N
,
5
2N
, . . . ,
(2N 1)
2N
and
t
i
= arccos
i
N

i=1
T
m
(t
i
)T
n
(t
i
) =
_

_
0 (m = n)
N/2 (m = n = 0)
N (m = n = 0)
The set of points t
i
are clearly the midpoints in of the rst case. The unequal spacing of
the points in x
i
(Nt
i
) compensates for the weight factor
W(x) = (1 x
2
)
1/2
in the continuous case.
9
Additional Identities of Chebyshev Polynomials
The Chebyshev polynomials are both orthogonal polynomials and the trigonometric cos nx
functions in disguise, therefore they satisfy a large number of useful relationships.
The dierentiation and integration properties are very important in analytical and numerical
work. We begin with
T
n+1
(x) = cos[(n + 1) cos
1
x]
and
T
n1
(x) = cos[(n 1) cos
1
x]
Dierentiating both expressions gives
1
(n + 1)
d[T
n+1
(x)]
dx
=
sin[(n + 1) cos
1
x

1 x
2
and
1
(n 1)
d[T
n1
(x)]
dx
=
sin[(n 1) cos
1
x

1 x
2
Subtracting the last two expressions yields
1
(n + 1)
d[T
n+1
(x)]
dx

1
(n 1)
d[T
n1
(x)]
dx
=
sin(n + 1) sin(n 1)
sin
or
T

n+1
(x)
(n + 1)

n1
(x)
(n 1)
=
2 cos n sin
sin
= 2T
n
(x) n 2
10
Therefore
T

2
(x) = 4T
1
T

1
(x) = T
0
T

0
(x) = 0
We have the formulas for the dierentiation of Chebyshev polynomials, therefore these for-
mulas can be used to develop integration for the Chebyshev polynomials:
_
T
n
(x)dx =
1
2
_
T
n+1
(x)
(n + 1)

T
n1
(x)
(n 1)
_
+ C n 2
_
T
1
(x)dx =
1
4
T
2
(x) + C
_
T
0
(x)dx = T
1
(x) + C
The Shifted Chebyshev Polynomials
For analytical and numerical work it is often convenient to use the half interval 0 x 1
instead of the full interval 1 x 1. For this purpose the shifted Chebyshev polynomials
are dened:
T

n
(x) = T
n
(2x 1)
Thus we have for the rst few polynomials
T

0
= 1
T

1
= 2x 1
T

2
= 8x
2
8x + 1
T

3
= 32x
3
48x
2
+ 18x 1
T

4
= 128x
4
256x
3
+ 160x
2
32x + 1
11
and the following powers of x as functions of T

n
(x);
1 = T

0
x =
1
2
(T

0
+ T

1
)
x
2
=
1
8
(3T

0
+ 4T

1
+ T

2
)
x
3
=
1
32
(10T

0
+ 15T

1
+ 6T

2
+ T

3
)
x
4
=
1
128
(35T

0
+ 56T

1
+ 28T

2
+ 8T

3
+ T

4
)
The recurrence relationship for the shifted polynomials is:
T

n+1
(x) = (4x 2)T

n
(x) T

n1
(x) T

0
(x) = 1
or
xT

n
(x) =
1
4
T

n+1
(x) +
1
2
T

n
(x) +
1
4
T

n1
(x)
where
T

n
(x) = cos
_
ncos
1
(2x 1)

= T
n
(2x 1)
Expansion of x
n
in a Series of T
n
(x)
A method of expanding x
n
in a series of Chebyshev polynomials employes the recurrence
relation written as
xT
n
(x) =
1
2
[T
n+1
(x) + T
n1
(x)] n = 1, 2, 3 . . .
xT
0
(x) = T
1
(x)
12
To illustrate the method, consider x
4
x
4
= x
2
(xT
1
) =
x
2
2
[T
2
+ T
0
] =
x
4
[T
1
+ T
3
+ 2T
1
]
=
1
4
[3xT
1
+ xT
3
] =
1
8
[3T
0
+ 3T
2
+ T
4
+ T
2
]
=
1
8
T
4
+
1
2
T
2
+
3
8
T
0
This result is consistent with the expansion of x
4
given in Table 2.
Approximation of Functions by Chebyshev Polynomials
Sometimes when a function f(x) is to be approximated by a polynomial of the form
f(x) =

n=0
a
n
x
n
+ E
N
(x) |x| 1
where |E
n
(x)| does not exceed an allowed limit, it is possible to reduce the degree of the
polynomial by a process called economization of power series. The procedure is to convert
the polynomial to a linear combination of Chebyshev polynomials:
N

n=0
a
n
x
n
=
N

n=0
b
n
T
n
(x) n = 0, 1, 2, . . .
It may be possible to drop some of the last terms without permitting the error to exceed
the prescribed limit. Since |T
n
(x)| 1, the number of terms which can be omitted is
determined by the magnitude of the coecient b.
The Chebyshev polynomials are useful in numerical work for the interval 1 x 1
because
1. |T
n
(x)] 1 within 1 x 1
2. The maxima and minima are of comparable value.
13
3. The maxima and minima are spread reasonably uniformly over the interval
1 x 1
4. All Chebyshev polynomials satisfy a three term recurrence relation.
5. They are easy to compute and to convert to and from a power series form.
These properties together produce an approximating polynomial which minimizes error in
its application. This is dierent from the least squares approximation where the sum of
the squares of the errors is minimized; the maximum error itself can be quite large. In
the Chebyshev approximation, the average error can be large but the maximum error is
minimized. Chebyshev approximations of a function are sometimes said to be mini-max
approximations of the function.
The following table gives the Chebyshev polynomial approximation of several power series.
14
Table 3: Power Series and its Chebyshev Approximation
1. f(x) = a
0
f(x) = a
0
T
0
2. f(x) = a
0
+ a
1
x
f(x) = a
0
T
0
+ a
1
T
1
3. f(x) = a
0
+ a
1
x + a
2
x
2
f(x) =
_
a
0
+
a
2
2
_
T
0
+ a
1
T
1
+
_
a
2
2
_
T
2
4. f(x) = a
0
+ a
1
x + a
2
x
2
+ a
3
x
3
f(x) =
_
a
0
+
a
2
2
_
T
0
+
_
a
1
+
3a
3
4
_
T
1
+
_
a
2
2
_
T
2
+
_
a
3
4
_
T
3
5. f(x) = a
0
+ a
1
x + a
2
x
2
+ a
3
x
3
+ a
4
x
4
f(x) =
_
a
0
+
a
2
2
+
a
3
8
_
T
0
+
_
a
1
+
3a
3
4
_
T
1
+
_
a
2
2
+
a
4
2
_
T
2
+
_
a
3
8
_
T
3
+
_
a
4
8
_
T
4
6. f(x) = a
0
+ a
1
x + a
2
x
2
+ a
3
x
3
+ a
4
x
4
+ a
5
x
5
f(x) =
_
a
0
+
a
2
2
+
3a
4
8
_
T
0
+
_
a
1
+
3a
3
4
+
5a
5
8
_
T
1
+
_
a
2
2
+
a
4
2
_
T
2
+
_
a
3
4
+
5a
5
16
_
T
3
+
_
a
4
8
_
T
4
+
_
a
5
16
_
T
5
15
Table 4: Formulas for Economization of Power Series
x = T
1
x
2
=
1
2
(1 + T
2
)
x
3
=
1
4
(3x + T
3
)
x
4
=
1
8
(8x
2
1 + T
4
)
x
5
=
1
16
(20x
3
5x + T
5
)
x
6
=
1
32
(48x
4
18x
2
+ 1 + T
6
)
x
7
=
1
64
(112x
5
56x
3
+ 7x + T
7
)
x
8
=
1
128
(256x
6
160x
4
+ 32x
2
1 + T
8
)
x
9
=
1
256
(576x
7
432x
5
+ 120x
3
9x + T
9
)
x
10
=
1
512
(1280x
8
1120x
6
+ 400x
4
50x
2
+ 1 + T
10
)
x
11
=
1
1024
(2816x
9
2816x
7
+ 1232x
5
220x
3
+ 11x + T
11
)
For easy reference the formulas for economization of power series in terms of Chebyshev are
given in Table 4.
16
Assigned Problems
Problem Set for Chebyshev Polynomials
1. Obtain the rst three Chebyshev polynomials T
0
(x), T
1
(x) and T
2
(x) by means of
the Rodrigues formula.
2. Show that the Chebyshev polynomial T
3
(x) is a solution of Chebyshevs equation of
order 3.
3. By means of the recurrence formula obtain Chebyshev polynomials T
2
(x) and T
3
(x)
given T
0
(x) and T
1
(x).
4. Show that T
n
(1) = 1 and T
n
(1) = (1)
n
5. Show that T
n
(0) = 0 if n is odd and (1)
n/2
if n is even.
6. Setting x = cos show that
T
n
(x) =
1
2
__
x + i
_
1 x
2
_
n
+
_
x i
_
1 x
2
_
n
_
where i =

1.
7. Find the general solution of Chebyshevs equation for n = 0.
8. Obtain a series expansion for f(x) = x
2
in terms of Chebyshev polynomials T
n
(x),
x
2
=
3

n=0
A
n
T
n
(x)
9. Express x
4
as a sum of Chebyshev polynomials of the rst kind.
17