Power Tower
The power tower of order
is defined as
|
(1)
|
where is Knuth
up-arrow notation (Knuth 1976), which in turn is defined by
|
(2)
|
together with
|
(3)
| |||
|
(4)
|
Rucker (1995, p. 74) uses the notation
|
(5)
|
and refers to this operation as "tetration."
A power tower can be implemented in the Wolfram Language as
PowerTower[a_, k_Integer] := Nest[Power[a, #]&, 1, k]
or
PowerTower[a_, k_Integer] := Power @@ Table[a, {k}]
The following table gives values of for
, 2, ... for small
.
| OEIS | ||
| 1 | A000027 | 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, ... |
| 2 | A000312 | 1, 4, 27, 256, 3125, 46656, ... |
| 3 | A002488 | 1, 16, 7625597484987, ... |
| 4 | 1, 65536, ... |
The following table gives
for
, 2, ... for small
.
| OEIS | ||
| 1 | A000012 | 1, 1, 1, 1, 1, 1, ... |
| 2 | A014221 | 2,
4, 16, 65536, |
| 3 | A014222 | 3, 27, 7625597484987, ... |
| 4 | 4, 256, |
Consider
and let
be defined as
|
(6)
|
(Galidakis 2004). Then for ,
is entire with series expansion:
|
(7)
|
Similarly, for ,
is analytic for
in the domain of the principal branch of
, with series expansion:
|
(8)
|
For , and
,
|
(9)
|
For , and
, and
|
(10)
|
The value of the infinite power tower , where
is an abbreviation for
, can be computed analytically by writing
|
(11)
|
taking the logarithm of both sides and plugging back in to obtain
|
(12)
|
Solving for
gives
|
(13)
|
where
is the Lambert W-function (Corless et al.
1996).
converges iff
(
; OEIS A073230
and A073229), as shown by Euler (1783) and
Eisenstein (1844) (Le Lionnais 1983; Wells 1986, p. 35).
Knoebel (1981) gave the following series for
|
(14)
| |||
|
(15)
|
(Vardi 1991).
The special value
is given by
|
(16)
| |||
|
(17)
| |||
|
(18)
|
(OEIS A077589 and A077590; Macintyre 1966).
The related function
|
(19)
|
converges only for ,
that is,
(OEIS A072364). The value it converges to is
the inverse of
which can be found by taking the logarithm of both sides of (19),
|
(20)
|
rearranging to
|
(21)
|
and then substituting to obtain
|
(22)
|
Solving the resulting equation for then gives the partial solution
|
(23)
|
which is valid for
(i.e.,
;
OEIS A072364 and A073226).
Taking
then gives
,
where
is the omega constant.
A continued fraction due to Khovanskii (1963) for the single iteration of is given by
|
(24)
|
The function
is plotted above along the real line and in the complex plane. It has series
expansion
|
(25)
|
(Trott 2004, p. 59). It has a minimum where
|
(26)
|
which has solution .
At this point, the function takes on the value
.
|
(27)
|
cannot be expressed in terms of a finite number of elementary functions, but some interesting definite integrals
of are
|
(28)
| |||
|
(29)
| |||
|
(30)
| |||
|
(31)
|
(OEIS A083648 and A073009; Spiegel 1968; Abramowitz and Stegun 1972; Havil 2003, pp. 44-45; Borwein et al. 2004, p. 5). Borwein et al. (2004, pp. 5 and 44) call these two integrals "a sophomore's dream."
The function
is plotted above along the real line and in the complex plane, where it shows beautiful
structure.
See also
Ackermann Function, Exponential Factorial, Exponential Function, Fermat Number, Joyce Sequence, Knuth Up-Arrow Notation, Lambert W-Function, Mills' Constant, MRB Constant, Nested Radical, Omega Constant, Power, Sophomore's Dream, Steiner's ProblemPortions of this entry contributed by Ioannis Galidakis
Explore with Wolfram|Alpha
References
Abramowitz, M. and Stegun, I. A. (Eds.). Handbook of Mathematical Functions with Formulas, Graphs, and Mathematical Tables, 9th printing. New York: Dover, 1972.Ash, J. M. "The Limit ofReferenced on Wolfram|Alpha
Power TowerCite this as:
Galidakis, Ioannis and Weisstein, Eric W. "Power Tower." From MathWorld--A Wolfram Resource. https://mathworld.wolfram.com/PowerTower.html