PFR

1 Applications of Jensen’s inequality

In this chapter, \(h\) denotes the function \(h(x) := x \log \frac{1}{x}\) for \(x \in [0,1]\).

Lemma 1.1 Concavity
#

\(h\) is strictly concave on \([0,\infty )\).

Proof

Check that \(h'\) is strictly monotone decreasing.

Lemma 1.2 Jensen
#

If \(S\) is a finite set, and \(\sum _{s \in S} w_s = 1\) for some non-negative \(w_s\), and \(p_s \in [0,1]\) for all \(s \in S\), then

\[ \sum _{s \in S} w_s h(p_s) \leq h(\sum _{s \in S} w_s p_s). \]
Proof

Apply Jensen and Lemma 1.1.

Lemma 1.3 Converse Jensen
#

If equality holds in the above lemma, then \(p_s = \sum _{s \in S} w_s h(p_s)\) whenever \(w_s \neq 0\).

Proof

Need some converse form of Jensen, not sure if it is already in Mathlib. May also wish to state it as an if and only if.

Lemma 1.4 log sum inequality
#

If \(S\) is a finite set, and \(a_s,b_s\) are non-negative for \(s\in S\), then

\[ \sum _{s\in S} a_s \log \frac{a_s}{b_s}\ge \left(\sum _{s\in S}a_s\right)\log \frac{\sum _{s\in S} a_s}{\sum _{s\in S} b_s}, \]

with the convention \(0\log \frac{0}{b}=0\) for any \(b\ge 0\) and \(0\log \frac{a}{0}=\infty \) for any \(a{\gt}0\).

Proof

Let \(B:=\sum _{s\in S} b_s\). Apply Jensen and Lemma 1.1 to show that \(\sum _{s\in S} \frac{b_s}{B} h(\frac{a_s}{b_s})\ge h(\frac{\sum _{s\in S} a_s}{B})\).

Lemma 1.5 converse log sum
#

If equality holds in Lemma 1.4, then \(a_s=r\cdot b_s\) for every \(s\in S\), for some constant \(r\in \mathbb {R}\).

Proof

By the fact that \(h\) is strictly concave and the equality condition of Jensen.