The term "Lebesgue constants" refers to either of two commonly used sets. The first is related to approximation of functions via Fourier series, while the other arises in the computation of Lagrange interpolating polynomials.
Assume a function is integrable over the interval
and
is the
th partial sum of the Fourier
series of
, so that
|
(1)
| |||
|
(2)
|
and
|
(3)
|
If
|
(4)
|
for all , then
|
(5)
|
and
is the smallest possible constant for which this holds
for all continuous functions
. The first few values of
are
|
(6)
| |||
|
(7)
| |||
|
(8)
| |||
|
(9)
| |||
|
(10)
| |||
|
(11)
| |||
|
(12)
|
|
(13)
| |||
|
(14)
|
(Zygmund 1959) and integral formulas include
|
(15)
| |||
|
(16)
|
(Hardy 1942). For large ,
|
(17)
|
This result can be generalized for a -periodic function with
continuous derivatives,
where
is an integer, satisfying
|
(18)
|
for all . In this case,
|
(19)
|
where
|
(20)
|
(Kolmogorov 1935, Zygmund 1959).
Watson (1930) showed that
|
(21)
|
where
|
(22)
| |||
|
(23)
| |||
|
(24)
|
(OEIS A243277), where is the gamma function,
is the Dirichlet lambda function, and
is the Euler-Mascheroni constant.
For a positive integer and
distinct nodes
with
,
define the
th
Lebesgue constant for the Lagrange
interpolating polynomial by
|
(25)
|
It is then true that
|
(26)
|
The efficiency of Lagrange interpolating polynomial approximation is related to the rate at which increases. Erdős (1961) proved that there exists
a positive constant such
that
|
(27)
|
for all . For the nodes given by the roots
of the Chebyshev polynomial of
the first kind
, Erdős (1961) notes that
|
(28)
|
where
is a positive constant
independent of
. Thus the leading coefficient
of
in (27) cannot be increased.
A node set is called canonical when it contains both endpoints
and 1. For
, there is a unique canonical
set minimizing
. It is symmetric
about 0, and its Lebesgue function attains the same
maximum on each of the
intervals between successive
nodes. These Bernstein-Erdős theorems were established by Kilgore (1977, 1978)
and de Boor and Pinkus (1978). More precisely, let
and
be the smallest and largest of those interval
maxima, and let
be the optimal value. Then
|
(29)
|
with both inequalities strict for every other canonical set.
The endpoint condition is essential to the uniqueness assertion. Without it, every three-node set with
has the minimal value
. Indeed, writing
, its Lebesgue function
is
for
and
for
, so
|
(30)
|