Zaslavsky G M, Edelman M
Courant Institute of Mathematical Sciences, New York University, 251 Mercer Street, New York, New York 10012, USA.
Phys Rev E Stat Nonlin Soft Matter Phys. 2005 Sep;72(3 Pt 2):036204. doi: 10.1103/PhysRevE.72.036204. Epub 2005 Sep 8.
Hamiltonian chaotic dynamics is, in general, not ergodic and the boundaries of the ergodic or quasiergodic area (stochastic sea, stochastic layers, stochastic webs, etc.) are sticky, i.e., trajectories can spend an arbitrarily long time in the vicinity of the boundaries with a nonexponentially small probability. Segments of trajectories imposed by the stickiness are called flights. The flights have polynomial dispersion that can lead to non-Gaussian statistics of displacements and to anomalous transport in phase space. In particular, the presence of flights influences the distribution of Poincaré recurrences. We use the distribution function of (l,t;epsilon, epsilon0) -separation of trajectories that at time instant t and trajectory length l are separated for the first time by epsilon<<1, being initially at a distance epsilon0 <<epsilon. The connection of this function, called the complexity function [Afraimovich and Zaslavsky, Chaos 13, 519 (2003)], with the distribution of Poincaré recurrences is established for three cases: (i) for the case of superdiffusion in standard and web maps for which the stickiness is defined by the boundaries of hierarchical sets of islands; (ii) for the case of the Sinai billiard with infinite horizon, where the stickiness is defined by zero-measure slits in the phase space; (iii) for the square billiard with a slit (bar-in-square billiard) where the Lyapunov exponent is zero and the stickiness is defined by the vicinity of the trajectory to the closest periodic trajectories obtained from the Diophantine approximation. Finally, the powerwise asymptotics of the Poincaré recurrences can be connected, in some cases, to the anomalous transport exponent.
一般而言,哈密顿混沌动力学不是遍历的,遍历或准遍历区域(随机海、随机层、随机网等)的边界是粘性的,即轨迹可以以非指数级小的概率在边界附近花费任意长的时间。由粘性导致的轨迹段称为飞行。飞行具有多项式扩散,这可能导致位移的非高斯统计以及相空间中的反常输运。特别地,飞行的存在会影响庞加莱回归的分布。我们使用轨迹的(l,t;ε,ε0) - 分离的分布函数,其中轨迹在时刻t和长度l时首次被ε << 1分开,初始时距离为ε0 << ε。这个被称为复杂度函数的函数([阿夫拉伊莫维奇和扎斯拉夫斯基,《混沌》13, 519 (2003)])与庞加莱回归分布之间的联系在三种情况下得以建立:(i) 对于标准映射和蛛网映射中的超扩散情况,其粘性由分层岛集的边界定义;(ii) 对于具有无限视界的 Sinai 台球情况,其粘性由相空间中的零测度狭缝定义;(iii) 对于带有狭缝的方台球(方内杆台球)情况,其李雅普诺夫指数为零,粘性由轨迹接近从丢番图逼近得到的最接近周期轨迹来定义。最后,在某些情况下,庞加莱回归的幂律渐近性可以与反常输运指数联系起来。