Ensemble averages and nonextensivity at the edge of chaos of one-dimensional maps

TL;DR

This study investigates ensemble averages of sensitivity and entropy production in one-dimensional maps, revealing a generalized Pesin-like relation and identifying a special q-value at the edge of chaos.

Contribution

It introduces a new family of dissipative maps and demonstrates a linear relation between generalized sensitivity and entropy, extending Pesin's theorem to nonextensive regimes.

Findings

Linear increase of generalized sensitivity and entropy with time at a specific q-value.

Coincidence of slopes for sensitivity and entropy, extending Pesin's theorem.

At the edge of chaos, the q-value differs from 1, indicating nonextensive behavior.

Abstract

Ensemble averages of the sensitivity to initial conditions $\xi(t)$ and the entropy production per unit time of a {\it new} family of one-dimensional dissipative maps, $x_{t+1}=1-ae^{-1/|x_t|^z}(z>0)$, and of the known logistic-like maps, $x_{t+1}=1-a|x_t|^z(z>1)$, are numerically studied, both for {\it strong} (Lyapunov exponent $\lambda_1>0$) and {\it weak} (chaos threshold, i.e., $\lambda_1=0$) chaotic cases. In all cases we verify that (i) both $<\ln_q \xi > [\ln_q x \equiv (x^{1-q}-1)/(1-q); \ln_1 x=\ln x]$ and $<S_q > [S_q \equiv (1-\sum_i p_i^q)/(q-1); S_1=-\sum_i p_i \ln p_i]$ {\it linearly} increase with time for (and only for) a special value of $q$, $q_{sen}^{av}$, and (ii) the {\it slope} of $<\ln_q \xi>$ and that of $<S_q>$ {\it coincide}, thus interestingly extending the well known Pesin theorem. For strong chaos, $q_{sen}^{av}=1$, whereas at the edge of chaos, $q_{sen}^{av}(z)<1$.