Approximate Gumbel Last Passage Percolation

3 Gumbel and Exponential Distributions

3.1 The Gumbel Distribution

Definition 8 Gumbel CDF
#

The Gumbel cumulative distribution function is:

\[ F_{\text{Gumbel}}(x) = \exp (-e^{-x}) \]
Lemma 9 Gumbel CDF is Continuous
#

The Gumbel CDF is continuous on \(\mathbb {R}\).

Definition 10 Gumbel Grid
#

A random field \(Y : \text{Edge} \to \Omega \to \mathbb {R}\) is a Gumbel grid if:

  1. The random variables \(\{ Y_e\} _{e \in \text{Edge}}\) are independent, and

  2. For each edge \(e\) and \(x \in \mathbb {R}\), \(\mathbb {P}(Y_e \leq x) = F_{\text{Gumbel}}(x)\).

3.2 Transformation to Exponential Distribution

Lemma 11 Gumbel Measure of Singletons is Zero
#

If \(Y\) is a Gumbel random variable, then for any \(y \in \mathbb {R}\):

\[ \mathbb {P}(Y = y) = 0 \]
Proof

This follows from the continuity of the Gumbel CDF. For any continuous CDF \(F\), the probability of a singleton is the difference \(F(y) - F(y^-)\), which is zero when \(F\) is continuous.

Lemma 12 Gumbel Probability of Complement
#

If \(Y\) is a Gumbel random variable, then:

\[ \mathbb {P}(Y \geq y) = 1 - F_{\text{Gumbel}}(y) \]
Lemma 13 Gumbel to Exponential Transformation
#

If \(Y\) is a Gumbel random variable, then \(\exp (-Y)\) has the exponential distribution with rate 1. Specifically, for \(x \geq 0\):

\[ \mathbb {P}(\exp (-Y) \leq x) = 1 - e^{-x} \]
Proof

For \(x {\gt} 0\), we have:

\begin{align*} \mathbb {P}(\exp (-Y) \leq x) & = \mathbb {P}(-Y \leq \log x) \\ & = \mathbb {P}(Y \geq -\log x) \\ & = 1 - F_{\text{Gumbel}}(-\log x) \\ & = 1 - \exp (-\exp (\log x)) \\ & = 1 - e^{-x} \end{align*}
Lemma 14 Exponential Grid from Gumbel
#

If \(Y\) is a Gumbel grid, then \(E_e = \exp (-Y_e)\) forms a grid of independent exponential random variables with rate 1.

Proof

The independence follows from the fact that \(\exp \) is a measurable function and composition with independent random variables preserves independence. The CDF property follows from the previous lemma applied to each edge.