arXiv subjects
Publications and source records attributed to Paolo Grigolini.
This paper presents a wavelet multiresolution analysis of a time series dataset to study the correlation between the real temperature data and three temperature model reconstructions at different scales. We show that the Mann et.al. model reconstructs the temperature better at all temporal resolutions. We show and discuss the wavelet multiresolution analysis of the Mann's temperature reconstruction for the period from 1400 to 2000 A.D.E.
The methods currently used to determine the scaling exponent of a complex dynamic process described by a time series are based on the numerical evaluation of variance. This means that all of them can be safely applied only to the case where ordinary statistical properties hold true even if strange kinetics are involved. We illustrate a method of statistical analysis based on the Shannon entropy of the diffusion process generated by the time series, called Diffusion Entropy Analysis (DEA). We adopt artificial Gauss and Lévy time series, as prototypes of ordinary and anomalus statistics, respectively, and we analyse them with the DEA and four ordinary methods of analysis, some of which are very popular. We show that the DEA determines the correct scaling exponent even when the statistical properties, as well as the dynamic properties, are anomalous. The other four methods produce correct results in the Gauss case but fail to detect the correct scaling in the case of Lévy statistics.
We study the dynamic and thermodynamic origin of non-canonical equilibria, and we discuss their connection with the generalized central limit theorem and the micro-canonical Boltzmann principle. We reach the conclusion that the zeroth law of thermodynamics and the Boltzmann principle are fulfilled thanks to an apparent fault turned into a benefit: the dynamic approach can only produce a truncated form of inverse power law equilibrium.
We analyze the waiting time distribution of time distances $τ$ between two nearest-neighbor flares. This analysis is based on the joint use of two distinct techniques. The first is the direct evaluation of the distribution function $ψ(τ)$, or of the probability, $Ψ(tau)$, that no time distance smaller than a given $τ$ is found. We adopt the paradigm of the inverse power law behavior, and we focus on the determination of the inverse power index $μ$, without ruling out different asymptotic properties that might be revealed, at larger scales, with the help of richer statistics. The second technique, called Diffusion Entropy (DE) method, rests on the evaluation of the entropy of the diffusion process generated by the time series. The details of the diffusion process depend on three different walking rules, which determine the form and the time duration of the transition to the scaling regime, as well as the scaling parameter $δ$. With the first two rules the information contained in the time series is transmitted, to a great extent, to the transition, as well as to the scaling regime. The same information is essentially conveyed, by using the third rules, into the scaling regime, which, in fact, emerges very quickly after a fast transition process. We show that the significant information hidden within the time series concerns memory induced by the solar cycle, as well as the power index $μ$. The scaling parameter $δ$ becomes a simple function of $μ$, when memory is annihilated. Thus, the three walking rules yield a unique and precise value of $μ$ if the memory is wisely taken under control, or cancelled by shuffling the data. All this makes compelling the conclusion that $μ= 2.138 \pm 0.01$.
We propose a new method of statistical analysis of nucleotide sequences yielding the true scaling without requiring any form of de-trending. With the help of artificial sequences that are proved to be statistically equivalent to the real DNA sequences we find that power-law correlations are present in both coding and non-coding sequences, in accordance with the recent work of other authors. We also afford a compelling evidence that these long-range correlations generate Lévy statistics in both types of sequences.
We study time series concerning rare events. The occurrence of a rare event is depicted as a jump of constant intensity always occurring in the same direction, thereby generating an asymmetric diffusion process. We consider the case where the waiting time distribution is an inverse power law with index $μ$. We focus our attention on $μ<3$, and we evaluate the scaling $δ$ of the resulting diffusion process. We prove that $δ$ gets its maximum value, $δ=1$, corresponding to the ballistic motion, at $μ=2$. We study the resulting diffusion process by means of joint use of the continuous time random walk and of the generalized central limit theorem, as well as adopting numerical treatment. We show that rendering asymmetric the diffusion process yelds the significant benefit of enhancing the value of the scaling parameter $δ$. Furthermore, this scaling parameter becomes sensitive to the power index $μ$ in the whole region $1<μ<3$. Finally, we show our method in action on real data concerning human heartbeat sequences.
Entanglement is the fundamental quantum property behind the now popular field of quantum transport of information. This quantum property is incompatible with the separation of a single system into two uncorrelated subsystems. Consequently, it does not require the use of an additive form of entropy. We discuss the problem of the choice of the most convenient entropy indicator, focusing our attention on a system of 2 qubits, and on a special set, denoted by $\Im$. This set contains both the maximally and the partially entangled states that are described by density matrices diagonal in the Bell basis set. We select this set for the main purpose of making more straightforward our work of analysis. As a matter of fact, we find that in general the conventional von Neumann entropy is not a monotonic function of the entanglement strength. This means that the von Neumann entropy is not a reliable indicator of the departure from the condition of maximum entanglement. We study the behavior of a form of non-additive entropy, made popular by the 1988 work by Tsallis. We show that in the set $\Im$, implying the key condition of non-vanishing entanglement, this non-additive entropy indicator turns out to be a strictly monotonic function of the strength of the entanglement, if entropy indexes $q$ larger than a critical value $Q$ are adopted. We argue that this might be a consequence of the non-additive nature of the Tsallis entropy, implying that the world is quantum and that uncorrelated subsystems do not exist.
We study a spatial diffusion process generated by velocity fluctuations of intermittent nature. We note that intermittence reduces the entropy production rate while enhancing the diffusion strength. We study a case of space-dependent intermittence and prove it to result in a deviation from uniform distribution. This macroscopic effect can be used to measure the relative value of the trajectory entropy.
We argue that a process of social interest is a balance of order and randomness, thereby producing a departure from a stationary diffusion process. The strength of this effect vanishes if the order to randomness intensity ratio vanishes, and this property allows us to reveal, although in an indirect way, the existence of a finite order to randomness intensity ratio. We aim at detecting this effect. We introduce a method of statistical analysis alternative to the compression procedures, with which the limitations of the traditional Kolmogorov-Sinai approach are bypassed. We prove that this method makes it possible for us to build up a memory detector, which signals the presence of even very weak memory, provided that this is persistent over large time intervals. We apply the analysis to the study of the teen birth phenomenon and we find that the unmarried teen births are a manifestation of a social process with a memory more intense than that of the married teens. We attempt to give a social interpretation of this effect.
We study the time evolution of a quantum system without classical counterpart, undergoing a process of entropy increase due to the environment influence. We show that if the environment-induced decoherence is interpreted in terms of wave-function collapses, a symbolic sequence can be generated. We prove that the Kolmogorov-Sinai entropy of this sequence coincides with rate of von Neumann entropy increase.
In the case of fully chaotic systems the distribution of the Poincare'recurrence times is an exponential whose decay rate is the Kolmogorov-Sinai(KS) entropy.We address the discussion of the same problem, the connection between dynamics and thermodynamics,in the case of sporadic randomness,using the Manneville map as a prototype of this class of processes. We explore the possibility of relating the distribution of Poincare' recurrence times to `thermodynamics',in the sense of the KS entropy,also in the case of an inverse power law. This is the dynamic property that Zaslavsly [Phys.Today(8), 39(1999)] finds to be responsible for a striking deviation from ordinary statistical mechanics under the form of Maxwell's Demon effect. We show that this way of estabi- lishing a connection between thermodynamics and dynamics is valid only in the case of strong chaos. In the case of sporadic randomness, resulting at long times in the Levy diffusion processes,the sensitivity to initial conditions is initially an inverse pow erlaw,but it becomes exponential in the long-time scale, whereas the distribution of Poincare times keeps its inverse power law forever. We show that a nonextensive thermodynamics would imply the Maxwell's Demon effect to be determined by memory and thus to be temporary,in conflict with the dynamic approach to Levy statistics. The adoption of heuristic arguments indicates that this effect,is possible, as a form of genuine equilibrium,after completion of the process of memory erasure.
We study sporadic randomness by means of a non-extensive form of Lyapunov coefficient. We recover from a different perspective the same conclusion as that of an earlier work, namely, that the ordinary Pesin theorem applies (P.Gaspard and X.-J. Wang, Proc. Natl. Acad. Sci. USA {\bf85}, 4591 (1988)). However, our theoretical analysis allows us to organize the numerical calculations so as to reveal the slow transition from a temporary form of non-extensive thermodynamics, corresponding to the prediction of a recent paper (M. Buiatti, P. Grigolini, A. Montagnini, Phys. Rev. Lett {\bf 82}, 3383 (1999)), to the ordinary extensive thermodynamics. We show that the transition takes place with a slow decay corresponding to the regression from a non-equilibrium initial condition to equilibrium condition.
We study the problem of entropy increase of the Bernoulli-shift map without recourse to the concept of trajectory and we discuss whether, and under which conditions if it does, the distribution density entropy coincides with the Kolmogorov-Sinai entropy, namely, with the trajectory entropy.
We consider the $d=1$ nonlinear Fokker-Planck-like equation with fractional derivatives $\frac{\partial}{\partial t}P(x,t)=D \frac{\partial^γ}{\partial x^γ}[P(x,t) ]^ν$. Exact time-dependent solutions are found for $ ν= \frac{2-γ}{1+ γ}$ ($-\infty<γ\leq 2$). By considering the long-distance {\it asymptotic} behavior of these solutions, a connection is established, namely $q=\frac{γ+3}{γ+1}$ ($0<γ\le 2$), with the solutions optimizing the nonextensive entropy characterized by index $q$ . Interestingly enough, this relation coincides with the one already known for Lévy-like superdiffusion (i.e., $ν=1$ and $0<γ\le 2$). Finally, for $(γ,ν)=(2, 0)$ we obtain $q=5/3$ which differs from the value $q=2$ corresponding to the $γ=2$ solutions available in the literature ($ν<1$ porous medium equation), thus exhibiting nonuniform convergence.
We focus our attention on dynamical processes characterized by an entropic index Q<1. According to the probabilistic arguments of Tsallis and Bukman [C.Tsallis, D.J. Bukman, Phys. Rev. E 54,R2197 (1996)] these processes are subdiffusional in nature. The non-extensive generalization of the Kolmogorov-Sinai entropy yielding the same entropic index implies the stationary condition. We note, on the other hand, that enforcing the stationary property on subdiffusion has the effect of producing a localization process occurring within a finite time scale. We thus conclude that the stationary dynamic processes with Q<1 must undergo a localization process occurring at finite time. We check the validity of this conclusion by means of a numerical treatment of the dynamics of the logistic map at the critical point.
The Levy diffusion processes are a form of non ordinary statistical mechanics resting, however, on the conventional Markov property. As a consequence of this, their dynamic derivation is possible provided that (i) a source of randomness is present in the corresponding microscopic dynamics and (ii) that the consequent process of memory erasure is properly taken into account by the theoretical treatment.
We study the entropy time evolution of a quantum mechanical model, which is frequently used as a prototype for Anderson's localization. Recently Latora and Baranger [V. Latora, M. Baranger, Phys. Rev.Lett. 82, 520(1999)] found that there exist three entropy regimes, a transient regime of passage from dynamics to thermodynamics, a linear in time regime of entropy increase, namely a thermodynamic regime of Kolmogorov kind, and a saturation regime. We use the non-extensive entropic indicator recently advocated by Tsallis [ C. Tsallis, J. Stat. Phys. 52, 479 (1988)] with a mobile entropic index q, and we find that with the adoption of the ``magic'' value q = Q = 1/2 the Kolmogorov regime becomes more extended and more distinct than with the traditional entropic index q = 1. We adopt a two-site model to explain these properties by means of an analytical treatment and we argue that Q =1/2 might be a typical signature of the occurrence of Anderson's localization.
We adapt the Kolmogorov-Sinai entropy to the non-extensive perspective recently advocated by Tsallis. The resulting expression is an average on the invariant distribution, which should be used to detect the genuine entropic index Q. We argue that the condition Q >1 is time dependent.