Difference between revisions of "Holley inequality"
(Importing text file) |
(Tex done) |
||
Line 1: | Line 1: | ||
− | An inequality for a finite [[ | + | An inequality for a finite [[distributive lattice]] $(\Gamma,{\prec})$, saying that if $\mu_1$ and $\mu_2$ map $\Gamma$ into $(0,\infty)$ and satisfy $\sum_\Gamma \mu_1(a) = \sum_\Gamma \mu_2(a)$ and |
+ | $$ | ||
+ | \mu_1(a) \mu_2(b) \le \mu_1(a \vee v) \mu_2(a \wedge b) | ||
+ | $$ | ||
+ | for all $a,b \in \Gamma$, then | ||
+ | $$ | ||
+ | \sum_\Gamma f(a) \mu_1(a) \ge \sum_\Gamma f(a) \mu_2(a) | ||
+ | $$ | ||
+ | for every $f : \Gamma \rightarrow \mathbf{R}$ that is non-decreasing in the sense that $a \prec b$ implies $f(a) \le f(b)$. It is due to R. Holley [[#References|[a4]]] and was motivated by the related [[FKG inequality]] [[#References|[a3]]]. It is an easy corollary [[#References|[a2]]] of the more powerful [[Ahlswede–Daykin inequality]] [[#References|[a1]]]. | ||
− | + | See also [[Correlation inequalities]]; [[Fishburn–Shepp inequality]]. | |
− | + | ====References==== | |
+ | <table> | ||
+ | <TR><TD valign="top">[a1]</TD> <TD valign="top"> R. Ahlswede, D.E. Daykin, "An inequality for the weights of two families, their unions and intersections" ''Z. Wahrscheinlichkeitsth. verw. Gebiete'' , '''43''' (1978) pp. 183–185</TD></TR> | ||
+ | <TR><TD valign="top">[a2]</TD> <TD valign="top"> P.C. Fishburn, "Correlation in partially ordered sets" ''Discrete Appl. Math.'' , '''39''' (1992) pp. 173–191</TD></TR> | ||
+ | <TR><TD valign="top">[a3]</TD> <TD valign="top"> C.M. Fortuin, P.N. Kasteleyn, J. Ginibre, "Correlation inequalities for some partially ordered sets" ''Comm. Math. Phys.'' , '''22''' (1971) pp. 89–103</TD></TR> | ||
+ | <TR><TD valign="top">[a4]</TD> <TD valign="top"> R. Holley, "Remarks on the FKG inequalities" ''Comm. Math. Phys.'' , '''36''' (1974) pp. 227–231</TD></TR> | ||
+ | </table> | ||
− | + | {{TEX|done}} | |
− | |||
− | |||
− | |||
− | |||
− | |||
− | |||
− |
Revision as of 17:56, 16 December 2016
An inequality for a finite distributive lattice $(\Gamma,{\prec})$, saying that if $\mu_1$ and $\mu_2$ map $\Gamma$ into $(0,\infty)$ and satisfy $\sum_\Gamma \mu_1(a) = \sum_\Gamma \mu_2(a)$ and $$ \mu_1(a) \mu_2(b) \le \mu_1(a \vee v) \mu_2(a \wedge b) $$ for all $a,b \in \Gamma$, then $$ \sum_\Gamma f(a) \mu_1(a) \ge \sum_\Gamma f(a) \mu_2(a) $$ for every $f : \Gamma \rightarrow \mathbf{R}$ that is non-decreasing in the sense that $a \prec b$ implies $f(a) \le f(b)$. It is due to R. Holley [a4] and was motivated by the related FKG inequality [a3]. It is an easy corollary [a2] of the more powerful Ahlswede–Daykin inequality [a1].
See also Correlation inequalities; Fishburn–Shepp inequality.
References
[a1] | R. Ahlswede, D.E. Daykin, "An inequality for the weights of two families, their unions and intersections" Z. Wahrscheinlichkeitsth. verw. Gebiete , 43 (1978) pp. 183–185 |
[a2] | P.C. Fishburn, "Correlation in partially ordered sets" Discrete Appl. Math. , 39 (1992) pp. 173–191 |
[a3] | C.M. Fortuin, P.N. Kasteleyn, J. Ginibre, "Correlation inequalities for some partially ordered sets" Comm. Math. Phys. , 22 (1971) pp. 89–103 |
[a4] | R. Holley, "Remarks on the FKG inequalities" Comm. Math. Phys. , 36 (1974) pp. 227–231 |
Holley inequality. Encyclopedia of Mathematics. URL: http://encyclopediaofmath.org/index.php?title=Holley_inequality&oldid=40021