Chaos, ergodicity, and the thermodynamics of lower-dimensional Hamiltonian systems

This paper uses the assumptions of ergodicity and a microcanonical distribution to compute estimates of the largest Lyapunov exponents in lower-dimensional Hamiltonian systems. That the resulting estimates are in reasonable agreement with the actual values computed numerically corroborates the intuition that chaos in such systems can be understood as arising generically from a parametric instability and that this instability can be modeled by a stochastic-oscillator equation (cf. Casetti, Clementi, and Pettini, Phys. Rev. E 54, 5969 (1996)), linearised perturbations of a chaotic orbit satisfying a harmonic-oscillator equation with a randomly varying frequency.