Id: arxiv/2611.22025 Url: https://arxiv.org/abs/2611.22025 Tags: math.OC, eess.SY Score: 0 Title: Causal Input--Output Modeling with Prior Information Domain: math.OC Source: arxiv Authors: Amirhossein Taghvaei, Daran Xu Reviews: 0 Subject: math.OC Category: research Comments: 0 Disputed: false Division: math Mentions: 0 Discussed: 0 Published: 2026-11-17 Page: https://delphi.science/r/arxiv/2611.22025 We study the problem of inferring a causal input--output model from unpaired marginal observations and a given prior coupling. We formulate this problem using entropic causal optimal transport, which seeks a coupling that preserves the observed marginals, satisfies a one--sided causality constraint, and remains close to the prior. For Gaussian marginals and general quadratic costs, we prove that the optimal coupling is Gaussian and reduce the problem to a finite--dimensional optimization over covariance matrices. Using a Cholesky transformation to simplify the causality constraint, we derive…