Search

Extended phase-space isokinetic methods in their deterministic [Minary et al., Phys. Rev. Lett. 93, 150201 (2004)] and stochastic forms [Leimkuhler et al., Mol. Phys. 111, 3579 (2013)] have proved tremendously successful in allowing multiple time-scale molecular dynamics simulations to be performed with very large time steps. These methods work by coupling the physical degrees of freedom to a set of Nosé-Hoover chain or Nosé-Hoover Langevin thermostats via an isokinetic constraint, which has the effect of avoiding resonance artifacts that plague multiple time-step algorithms. In this paper, we introduce a new resonance-free approach that achieves the same gains in time step but without the imposition of isokinetic constraints or the introduction of extended phase-space variables. Rather, we modify the physical Hamiltonian that effects the same regulation of resonances achieved by the isokinetic constraints. In so doing, we show that sampling errors can be controlled and performance improvements are possible within a simpler Hamiltonian framework. The method is demonstrated in simulations of the structure of liquid water and, in conjunction with enhanced sampling, in generation of the Ramachandran free-energy surface of the solvated alanine dipeptide. [ABSTRACT FROM AUTHOR]

Molecular dynamics is one of the most commonly used approaches for studying the dynamics and statistical distributions of physical, chemical, and biological systems using atomistic or coarse-grained models. It is often the case, however, that the interparticle forces drive motion on many time scales, and the efficiency of a calculation is limited by the choice of time step, which must be sufficiently small that the fastest force components are accurately integrated. Multiple time-stepping algorithms partially alleviate this inefficiency by assigning to each time scale an appropriately chosen step-size. As the fast forces are often computationally cheaper to evaluate than the slow forces, this results in a significant gain in efficiency. However, such approaches are limited by resonance phenomena, wherein motion on the fastest time scales limits the step sizes associated with slower time scales. In atomistic models of biomolecular systems, for example, resonances limit the largest time step to around 5-6 fs. Stochastic processes promote mixing and ergodicity in dynamical systems and reduce the impact of resonant modes. In this paper, we introduce a set of stochastic isokinetic equations of motion that are shown to be rigorously ergodic, largely free of resonances, and can be integrated using a multiple time-stepping algorithm which is easily implemented in existing molecular dynamics codes. The technique is applied to a simple, illustrative problem and then to a more realistic system, namely, a flexible water model. Using this approach outer time steps as large as 100 fs are shown to be possible. [ABSTRACT FROM AUTHOR]

Explicit reversible integrators, suitable for use in large-scale computer simulations, are derived for extended systems generating the canonical and isothermal-isobaric ensembles. The new methods are compared with the standard implicit (iterative) integrators on some illustrative example problems. In addition, modification of the proposed algorithms for multiple time step integration is outlined. [ABSTRACT FROM AUTHOR]