Large deviations for interacting multiscale particle systems (Q2105066)

The goal of this article is to obtain the large deviations principle (LDP) for interacting particle systems of diffusion type in multiscale environments. The authors study the combined effect of weak mean field interactions in a fast oscillating multiscale environment from the point of view of large deviations for the empirical measure of the particles. They use methods from weak convergence and stochastic control [\textit{P. Dupuis} and \textit{R. S. Ellis}, A weak convergence approach to the theory of large deviations. Chichester: John Wiley \& Sons (1997; Zbl 0904.60001)] and some methods for mean field stochastic control problems [\textit{K. Jańczak}, Probab. Math. Stat. 28, No. 1, 41--47 (2008; Zbl 1147.60036)]. Their main results are given in Theorem 3.7, that gives the large deviations principle of the empirical distribution of the particles in the combined limit \(N \to \infty\) and \(\varepsilon \downarrow 0 \). This result seems to extend some important results given in [\textit{D. A. Dawson} and \textit{J. Gärtner}, Stochastics 20, 247--308 (1987; Zbl 0613.60021); Dupuis and Ellis, loc. cit.; \textit{A. Budhiraja} et al., Ann. Probab. 40, No. 1, 74--102 (2012; Zbl 1242.60026)] or [\textit{M. Röckner} et al., Ann. Inst. Henri Poincaré, Probab. Stat. 57, No. 1, 547--576 (2021; Zbl 1491.60088)].