Infinite random power towers

Abstract: We prove a probabilistic generalization of the classic result that infinite power towers,

c^{c^{c d o t^{c d o t^{c} d o t}}}

, converge if and only if

c i n [e^{- e}, e^{1 / e}]

. Given an i.i.d. sequence

{A_{i}}_{i i n m a t h b b N}

, we find that convergence of the power tower

A_{1}^{A_{2}^{c d o t^{c d o t^{c} d o t}}}

is determined by the bounds of

A_{1}

's support,

a = m a t h r m i n f (m a t h r m s u p p (A_{1}))

and

b = m a t h r m i n f (m a t h r m s u p p (A_{1}))

. When

b i n [e^{- e}, e^{1 / e}]

,

a < 1 < b

, or

a = 0

, the power tower converges almost surely. When

b < e^{- e}

, we define a special function

G

such that almost sure convergence is equivalent to

a < G (b)

. Only in the case when

a = 1

and

b > e^{1 / e}

are the values of

a

and

b

insufficient to determine convergence. We show a rather complicated necessary and sufficient condition for convergence when

a = 1

and

b

is finite. We also briefly discuss the relationship between the distribution of

A_{1}

and the corresponding power tower

T = A_{1}^{A_{2}^{c d o t^{c d o t^{c} d o t s}}}

. For example, when

T s i m m a t h r m U n i f [0, 1]

, then the corresponding distribution of

A_{1}

is given by

U V

where

U, V s i m m a t h r m U n i f [0, 1]

are independent. We generalize this example by showing that for

and

r i n m a t h b b R

, there exists an i.i.d. sequence

{A_{i}}_{i i n m a t h b b N}

such that

U^{r} s t a c k r e l d = A_{1}^{A_{2}^{c d o t^{c d o t^{c} d o t}}}

if and only if

.

Recommendations

Cites work

This page was built for publication: Infinite random power towers