In mathematics , the Lehmer mean is a generalized mean named after Derrick Henry Lehmer .
definition
The Lehmer mean of positive real numbers for the level is defined as follows:
n
{\ displaystyle n}
a
1
,
...
,
a
n
{\ displaystyle a_ {1}, \ ldots, a_ {n}}
p
{\ displaystyle p}
L.
p
(
a
1
,
...
,
a
n
)
=
∑
k
=
1
n
a
k
p
∑
k
=
1
n
a
k
p
-
1
.
{\ displaystyle L_ {p} (a_ {1}, \ ldots, a_ {n}) = {\ frac {\ sum _ {k = 1} ^ {n} a_ {k} ^ {p}} {\ sum _ {k = 1} ^ {n} a_ {k} ^ {p-1}}}.}
There is also a form of the clay remedy with (positive) weights . The weighted Lehmer mean is:
w
=
(
w
1
,
...
,
w
n
)
{\ displaystyle w = (w_ {1}, \ ldots, w_ {n})}
L.
p
,
w
(
a
1
,
...
,
a
n
)
=
∑
k
=
1
n
w
k
a
k
p
∑
k
=
1
n
w
k
a
k
p
-
1
.
{\ displaystyle L_ {p, w} (a_ {1}, \ ldots, a_ {n}) = {\ frac {\ sum _ {k = 1} ^ {n} w_ {k} a_ {k} ^ { p}} {\ sum _ {k = 1} ^ {n} w_ {k} a_ {k} ^ {p-1}}}.}
properties
The following applies to the Lehmer remedy
lim
p
→
-
∞
L.
p
(
a
1
,
...
,
a
n
)
=
min
{
a
1
,
...
,
a
n
}
{\ displaystyle \ lim _ {p \ to - \ infty} L_ {p} (a_ {1}, \ ldots, a_ {n}) = \ min \ {a_ {1}, \ ldots, a_ {n} \ }}
is the minimum value.
L.
0
(
a
1
,
...
,
a
n
)
=
n
∑
k
=
1
n
1
a
k
=
n
1
a
1
+
⋯
+
1
a
n
{\ displaystyle L_ {0} (a_ {1}, \ ldots, a_ {n}) = {\ frac {n} {\ sum _ {k = 1} ^ {n} {\ frac {1} {a_ { k}}}}} = {\ frac {n} {{\ frac {1} {a_ {1}}} + \ cdots + {\ frac {1} {a_ {n}}}}}}
is the harmonic mean .
For is the geometric mean .
n
=
2
{\ displaystyle n = 2}
L.
1
2
(
a
1
,
a
2
)
=
a
1
+
a
2
1
a
1
+
1
a
2
=
a
1
a
2
{\ displaystyle L _ {\ frac {1} {2}} (a_ {1}, a_ {2}) = {\ frac {{\ sqrt {a_ {1}}} + {\ sqrt {a_ {2}} }} {{\ frac {1} {\ sqrt {a_ {1}}}} + {\ frac {1} {\ sqrt {a_ {2}}}}}} = {\ sqrt {a_ {1} a_ {2}}}}
L.
1
(
a
1
,
...
,
a
n
)
=
∑
k
=
1
n
a
k
n
=
a
1
+
⋯
+
a
n
n
{\ displaystyle L_ {1} (a_ {1}, \ ldots, a_ {n}) = {\ frac {\ sum _ {k = 1} ^ {n} a_ {k}} {n}} = {\ frac {a_ {1} + \ cdots + a_ {n}} {n}}}
is the arithmetic mean .
L.
2
(
a
1
,
...
,
a
n
)
=
∑
k
=
1
n
a
k
2
a
1
+
⋯
+
a
n
{\ displaystyle L_ {2} (a_ {1}, \ ldots, a_ {n}) = {\ frac {\ sum _ {k = 1} ^ {n} a_ {k} ^ {2}} {a_ { 1} + \ cdots + a_ {n}}}}
is the contrarian means already known to Eudoxus by Knidos .
lim
p
→
∞
L.
p
(
a
1
,
...
,
a
n
)
=
Max
{
a
1
,
...
,
a
n
}
{\ displaystyle \ lim _ {p \ to \ infty} L_ {p} (a_ {1}, \ ldots, a_ {n}) = \ max \ {a_ {1}, \ ldots, a_ {n} \} }
is the maximum value.
In contrast to the other five special cases, the contrarian means is not monotonic, ie from for all does not follow .
a
i
≤
b
i
{\ displaystyle a_ {i} \ leq b_ {i}}
i
{\ displaystyle i}
L.
2
(
a
1
,
...
,
a
n
)
≤
L.
2
(
b
1
,
...
,
b
n
)
{\ displaystyle L_ {2} (a_ {1}, \ ldots, a_ {n}) \ leq L_ {2} (b_ {1}, \ ldots, b_ {n})}
Individual evidence
↑ Hischer / Lambert: "What is a numerical mean?"
↑ Axioms of mean values
literature
DH Lehmer: On the compounding of certain means. J. Math. Anal. Appl. 36 (1971) pp. 183-200
P. S. Bullen: Handbook of Means and Their Inequalities . Kluwer Acad. Pub. 2003, ISBN 1-4020-1522-4 (comprehensive discussion of mean values and the inequalities associated with them).
<img src="https://de.wikipedia.org/wiki/Special:CentralAutoLogin/start?type=1x1" alt="" title="" width="1" height="1" style="border: none; position: absolute;">