Transient subtraction method: An alternative approach for variance reduction in dissipative particle dynamics
Abstract
The efficient and accurate computation of transport coefficients remains a numerical challenge in dissipative particle dynamics (DPD), which can require substantial sampling to control statistical uncertainty. We extend recent results obtained in Langevin dynamics to consider the transient subtraction method for reducing the variance of transport coefficient estimations in DPD. The system is initialised by perturbing equilibrium samples, and its relaxation towards equilibrium is monitored under the reference dynamics. A reference trajectory, initialised at equilibrium and driven by the same Brownian noise, provides a control variate for the transient response. We apply the transient subtraction method to the computation of the mobility and the shear viscosity in a Lennard-Jones fluid in DPD. The shear viscosity is obtained from the time-integrated response of a Fourier observable associated with the longitudinal velocity. Numerical comparisons with the naive transient estimator show that the transient subtraction method reduces the variance at various integration times, and the reduction is more pronounced with a higher friction coefficient. These results indicate that the transient subtraction method is able to improve the statistical precision of transport coefficient estimations in DPD and reduce the number of trajectories needed to attain a prescribed standard error.
Related
- Reduction of finite-size effects for second-order Møller-Plesset perturbation theory with singularity subtraction
- A variational framework for variance reduction in lattice field theory
- Problem-aware sampling for bias and variance reduction in randomised Hamiltonian simulation
- Structure-Aware Variance Reduction for Unbiased Randomized Hamiltonian Simulation
- TETRIS-Q: Tiling-based Effective Transient-fault Reduction on Interleaved Superconducting Qubits
BibTeX
@misc{wu2026transient,
title = {{Transient subtraction method: An alternative approach for variance reduction in dissipative particle dynamics}},
author = {Xinyi Wu and Xiaocheng Shang},
howpublished = {arXiv},
year = {2026},
month = nov,
doi = {10.48550/arxiv.2611.00330},
eprint = {2611.00330},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2611.00330},
note = {CC BY 4.0},
}
RIS
TY - UNPB
TI - Transient subtraction method: An alternative approach for variance reduction in dissipative particle dynamics
AU - Xinyi Wu
AU - Xiaocheng Shang
PY - 2026
DA - 2026/11/01
PB - arXiv
DO - 10.48550/arxiv.2611.00330
UR - https://arxiv.org/abs/2611.00330
AB - The efficient and accurate computation of transport coefficients remains a numerical challenge in dissipative particle dynamics (DPD), which can require substantial sampling to control statistical uncertainty. We extend recent results obtained in Langevin dynamics to consider the transient subtraction method for reducing the variance of transport coefficient estimations in DPD. The system is initialised by perturbing equilibrium samples, and its relaxation towards equilibrium is monitored under the reference dynamics. A reference trajectory, initialised at equilibrium and driven by the same Brownian noise, provides a control variate for the transient response. We apply the transient subtraction method to the computation of the mobility and the shear viscosity in a Lennard-Jones fluid in DPD. The shear viscosity is obtained from the time-integrated response of a Fourier observable associated with the longitudinal velocity. Numerical comparisons with the naive transient estimator show that the transient subtraction method reduces the variance at various integration times, and the reduction is more pronounced with a higher friction coefficient. These results indicate that the transient subtraction method is able to improve the statistical precision of transport coefficient estimations in DPD and reduce the number of trajectories needed to attain a prescribed standard error.
ER -