{"id":"arxiv/2607.04466","url":"https://arxiv.org/abs/2607.04466","tags":["math.ST","math.PR","stat.CO"],"score":0,"title":"On the mixing properties of some preconditioned multiproposal Markov Chain Monte Carlo algorithms","domain":"math.ST","source":"arxiv","authors":["Giulia Carigi","Nathan E. Glatt-Holtz","Cecilia F. Mondaini","Guillermina Senn"],"reviews":0,"subject":"math.ST","summary":"We study two recently discovered \"dimension-free\" Monte Carlo sampling algorithms, the multiproposal and multiple-try preconditioned Crank-Nicolson methods (mpCN and MTpCN). These methods were designed to address certain non-parametric (i.e. infinite-dimensional) sampling problems, defined relative to a Gaussian reference measure, by combining proposal and acceptance mechanisms that take non-trivial advantage of parallel computing architectures. We provide the first rigorous analysis of both algorithms, establishing exponential convergence to the target measure through the weak Harris…","category":"research","comments":0,"disputed":false,"division":"math","mentions":0,"discussed":0,"published":"2026-07-05","reactions":{}}