[Paper Review] Comment on "Entropy of Classical Systems with Long-Range Interactions" by T.M. Rocha Filho et al, PRL 95, 190601 (2005)

Andrea Rapisarda, Alessandro Pluchino+1 more|arXiv (Cornell University)|Jan 18, 2006

Advanced Thermodynamics and Statistical Mechanics3 citations

TL;DR

This comment challenges the conclusions of Rocha Filho et al. (2005) on entropy in long-range interacting systems, arguing that nonextensive statistical mechanics (Tsallis statistics) is not only valid but essential for describing quasi-stationary states in Hamiltonian systems like the HMF model. It demonstrates that generalized statistics quantitatively explains anomalous diffusion and q-exponential decay, invalidating the assumption of equiprobable microstates and standard Boltzmann-Gibbs entropy in such systems.

ABSTRACT

In a recent letter (PRL 95, 190601 (2005)), T.M. Rocha Filho and coworkers address the very interesting issue of the entropic form to be used for Hamiltonians with long-range interactions. In our opinion the letter misses several points quite debated in the recent literature which are of fundamental importance for a complete discussion of this problem. Moreover it contains several statements which are not true or not corroborated by any evidence. In this comment we discuss these arguments and severely question the generality of the conclusions of this letter.

Motivation & Objective

• To challenge the claim that nonextensive statistics is meaningless or limited in scope for long-range interacting systems.
• To correct inaccuracies in Rocha Filho et al.'s analysis, particularly regarding microstate equiprobability and entropy form.
• To emphasize the importance of nonextensive statistical mechanics in describing nonhomogeneous quasi-stationary states in long-range Hamiltonian systems.
• To highlight that finite-size effects and nonergodicity invalidate the general conclusions of the original paper.
• To demonstrate that generalized statistics successfully predicts q-exponential decay and anomalous diffusion in the HMF model across diverse initial conditions and system sizes.

Proposed method

• Critique of the original paper's assumptions, particularly the a-priori assumption of homogeneous quasi-stationary states and equiprobable microstates.
• Reference to numerical evidence from the HMF model showing strong hierarchical microscopic structures sensitive to initial conditions.
• Application of generalized statistics (Tsallis entropy) to predict q-exponential decay of velocity correlation functions.
• Use of the generalized diffusion coefficient formula γ = 2/(3−q) to quantitatively match observed anomalous diffusion.
• Citation of prior rebuttals to criticisms in Refs. [13,14] of the original paper, affirming the validity of nonextensive statistics.
• Comparison of standard Boltzmann-Gibbs entropy with generalized statistics under different limits (N→∞ before t→∞ vs. t→∞ before N→∞), showing the latter's superiority in nonstationary regimes.

Experimental results

Research questions

• RQ1 Is the use of nonextensive statistical mechanics invalid or limited in scope for long-range interacting systems, as claimed by Rocha Filho et al.?
• RQ2 Does the assumption of equiprobable microstates hold for quasi-stationary states in long-range Hamiltonian systems, especially in the thermodynamic limit?
• RQ3 Can standard Boltzmann-Gibbs entropy correctly describe the velocity distribution in long-range systems during the quasi-stationary state?
• RQ4 Is the generalized statistics framework capable of quantitatively predicting anomalous diffusion and velocity correlation decay in the HMF model?
• RQ5 How do finite-size effects and the order of limits (N→∞ vs. t→∞) impact the validity of standard statistical mechanics in these systems?

Key findings

• The claim that nonextensive statistics is meaningless or limited is incorrect; it has proven useful across diverse physical and non-physical applications.
• The assumption of equiprobable microstates is invalid for quasi-stationary states in long-range systems when t→∞ is taken after N→∞, indicating strong nonergodicity.
• Standard Boltzmann-Gibbs entropy is only correct in the limit where N→∞ is taken before t→∞, but fails when the time limit is taken first, as in physical quasi-stationary states.
• The generalized statistics framework successfully predicts the q-exponential decay of velocity correlation functions observed in the HMF model.
• The generalized diffusion coefficient γ = 2/(3−q) has been quantitatively verified in simulations across multiple system sizes and initial conditions.
• Finite-size effects and the hierarchical microscopic structure of quasi-stationary states invalidate the general conclusions of the original paper, which assumed homogeneity and equiprobability a priori.