A correlation inequality for random points in a hypercube with some implications

Publication date: 1 September 2022

Abstract: Let

p r e c

be the product order on

m a t h b b R^{k}

and assume that

X_{1}, X_{2}, l d o t s, X_{n}

(

n g e q 3

) are i.i.d. random vectors distributed uniformly in the unit hypercube

[0, 1]^{k}

. Let

S

be the (random) set of vectors in

m a t h b b R^{k}

that

p r e c

-dominate all vectors in

X_{3}, .., X_{n}

, and let

W

be the set of vectors that are not

p r e c

-dominated by any vector in

X_{3}, .., X_{n}

. The main result of this work is the correlation inequality �egin{equation*} P(X_2in W|X_1in W)leq P(X_2in W|X_1in S),. end{equation*} For every

1 l e q i l e q n

let

E_{i, n}

be the event that

X_{i}

is not

p r e c

-dominated by any of the other vectors in

X_{1}, l d o t s, X_{n}

. The main inequality yields an elementary proof for the result that the events

E_{1, n}

and

E_{2, n}

are asymptotically independent as

n o i n f t y

. Furthermore, we derive a related combinatorial formula for the variance of the sum

s u m_{i = 1}^{n} e x t b f 1_{E_{i, n}}

, i.e. the number of maxima under the product order

p r e c

, and show that certain linear functionals of partial sums of

e x t b f 1_{E_{i, n}}; 1 l e q i l e q n

are asymptotically normal as

n o i n f t y

.

This page was built for publication: A correlation inequality for random points in a hypercube with some implications