Problem
Research questions and friction points this paper is trying to address.
Develops parallel-in-time method for McKean-Vlasov SDEs
Uses Monte Carlo and moment ODEs as propagators
Handles both unimodal and bimodal particle distributions
Monte-Carlo/Moments micro-macro Parareal method for unimodal and bimodal scalar McKean-Vlasov SDEs
🤖 AI Summary
This work addresses the efficient numerical solution of unimodal and bimodal scalar McKean–Vlasov stochastic differential equations (SDEs). We propose a micro-macro parallel-in-time algorithm: at the fine level, particle-based Monte Carlo simulations capture the full distributional evolution; at the coarse level, ordinary differential equations (ODEs) governing the mean and variance accelerate iteration. For the bimodal case, we introduce a novel multi-region moment model that partitions phase space into locally unimodal subdomains—thereby overcoming the degeneracy of conventional coarse predictors under multimodality. The algorithm synergistically integrates Parareal-type parallel-in-time computation with adaptive moment modeling to ensure stable convergence from coarse moment approximations to the true distribution. Linear convergence is rigorously established for linear cases. Numerical experiments demonstrate favorable weak scalability, reduced iteration counts for bimodal problems, and significantly improved accuracy over standalone moment models.
📝 Abstract
We propose a micro-macro parallel-in-time Parareal method for scalar McKean-Vlasov stochastic differential equations (SDEs). In the algorithm, the fine Parareal propagator is a Monte Carlo simulation of an ensemble of particles, while an approximate ordinary differential equation (ODE) description of the mean and the variance of the particle distribution is used as a coarse Parareal propagator to achieve speedup. We analyse the convergence behaviour of our method for a linear problem and provide numerical experiments indicating the parallel weak scaling of the algorithm on a set of examples. We show, with numerical experiments, that convergence typically takes place in a low number of iterations, depending on the quality of the ODE predictor. For bimodal SDEs, we avoid quality deterioration of the coarse predictor (compared to unimodal SDEs) through the usage of multiple ODEs, each describing the mean and variance of the particle distribution in locally unimodal regions of the phase space. The benefit of the proposed algorithm can be viewed through two lenses: (i) through the parallel-in-time lens, speedup is obtained through the use of a very cheap coarse integrator (an ODE moment model), and (ii) through the moment models lens, accuracy is iteratively gained through the use of parallel machinery as a corrector. In contrast to the isolated use of a moment model, the proposed method (iteratively) converges to the true distribution generated by the SDE.
Innovation
Methods, ideas, or system contributions that make the work stand out.
Micro-macro Parareal method combines Monte Carlo and ODE propagators
Multiple ODEs model local unimodal regions for bimodal SDEs
Cheap ODE moment model enables parallel-in-time speedup
🔎 Similar Papers
No similar papers found.
