Abstract:

This lecture provides a formal definition of classical chaos, breaking it down into three core ingredients: transitivity (ergodicity), the existence of a dense set of unstable periodic orbits, and sensitive dependence on initial conditions. Using the solvable doubling map as a primary example, the lecture introduces key quantitative measures like topological entropy. It then explores the concepts of equilibrium and non-equilibrium dynamics, defining invariant measures and the process of mixing, which describes how a system approaches equilibrium, analyzed via the Perron-Frobenius operator. The discussion then transitions from simple 1D maps to more physically relevant Hamiltonian systems, introducing the stroboscopic map as a technique to analyze time-periodic systems. The lecture culminates with an in-depth look at the standard map, a paradigmatic model for 2D area-preserving maps derived from a "kicked rotor" system, visually demonstrating the transition from regular to chaotic motion as a control parameter is varied.

A Lecture on the Foundations of Classical Chaos

• 0:00:14 Defining Classical Chaos - Ingredient 1: Transitivity: A chaotic system must be transitive (or ergodic), meaning that almost all initial conditions will eventually come arbitrarily close to any other point in the allowed phase space. The doubling map (2x mod 1) is used as a key example.
• 0:03:26 Ingredient 2: Dense Periodic Orbits: Chaos requires a dense but measure-zero set of periodic orbits. This means that arbitrarily close to any point, there is a periodic point, yet the probability of randomly picking a periodic point is zero. For the doubling map, these correspond to repeating binary sequences.
• 0:07:02 Topological Entropy: The number of periodic orbits of period L grows exponentially. The rate of this exponential growth, h_T, is called the topological entropy and is a key characteristic of a chaotic system. For the doubling map, h_T = log(2).
• 0:19:37 Ingredient 3: Sensitive Dependence on Initial Conditions: This is the "butterfly effect," where nearby initial points separate over time. It is proven that for many systems, the first two ingredients (transitivity and dense periodic orbits) are sufficient to imply this third one.
• 0:25:19 Equilibrium and Invariant Measures: In a chaotic system, an orbit spends a certain fraction of its time in different regions of phase space. This leads to an equilibrium or "invariant" density. For the logistic map at r=4, this is the arcsin distribution; for the doubling map, it is a uniform distribution.
• 0:28:46 Birkhoff's Ergodic Theorem: For an ergodic system, the long-time average of an observable along a single trajectory equals the phase space average weighted by the invariant measure. This connects the dynamics of a single point to the statistical properties of the whole system.
• 0:36:44 Approach to Equilibrium (Mixing): A system is "mixing" if any initial non-equilibrium distribution of points evolves towards the invariant (equilibrium) distribution. This is a stronger condition than ergodicity. The evolution of these densities is governed by the Perron-Frobenius operator.
• 1:03:59 Chaos in Hamiltonian Systems: Chaos can arise in Hamiltonian systems with 1.5 degrees of freedom (one position, one momentum, plus periodic time-dependence), such as a forced pendulum or kicked rotor.
• 1:06:03 Stroboscopic Maps: To analyze these periodically driven systems, we use a stroboscopic map. We observe the system's state (position and momentum) only at integer multiples of the driving period, turning the continuous flow into a discrete 2D area-preserving map.
• 1:23:22 Kicked Systems and the Standard Map: A useful way to create analytically tractable chaotic maps is to use a Hamiltonian with "kicks" (delta functions in time). The "kicked rotor" gives rise to the famous standard map, a one-parameter family of maps that serves as a paradigm for studying chaos.
• 1:29:22 Phase Space of the Standard Map: As the single parameter K (kicking strength) is increased, the phase space of the standard map transitions from being regular and integrable (K=0) to a mixed phase space with islands of stability surrounded by a "chaotic sea," and finally to globally chaotic behavior where a single orbit appears to fill the entire space.