Adrien Cances

Adrien Cances, Quentin Mérigot, Luca Nenna

We study a nonlinear multimarginal optimal transport problem arising in risk management, where the objective is to maximize a spectral risk measure of the pushforward of a coupling by a cost function. Although this problem is inherently nonlinear, it is known to have an equivalent linear reformulation as a multimarginal transport problem with an additional marginal. We introduce a Lagrangian particle discretization of this problem, in which admissible couplings are approximated by uniformly weighted point clouds, and marginal constraints are enforced through Wasserstein penalization. We prove quantitative convergence results for this discretization as the number of particles tends to infinity. The convergence rate is shown to be governed by the uniform quantization error of an optimal solution, and can be bounded in terms of the geometric properties of its support, notably its box dimension. In the case of univariate marginals and supermodular cost functions, where optimal couplings are known to be comonotone, we obtain sharper convergence rates expressed in terms of the asymptotic quantization errors of the marginals themselves. We also discuss the particular case of conditional value at risk, for which the problem reduces to a multimarginal partial transport formulation. Finally, we illustrate our approach with numerical experiments in several application domains, including risk management and partial barycenters, as well as some artificial examples with a repulsive cost.

Adrien Cances, Luca Nenna, Daniyar Omarov et al.

We consider entropically regularized, semi-discrete versions of variational problems on the set of probability measures involving optimal transport as well as other terms. We prove that the solutions can be characterized by well-posed ordinary differential equations in the regularization parameter. The initial conditions for these equations, corresponding to solutions to completely regularized problems, are typically known explicitly. The ODE can then be solved to recover the solution for an arbitrary degree of regularization; we verify that the solution is continuous in the regularization parameter, implying that taking the limit of the trajectory yields the solution to the fully unregularized problem. We establish analogous results for a version of the problem when the non-optimal transport term is not scaled with the regularization parameter. We exploit our characterization to numerically solve several example problems using standard ODE methods; this strategy exhibits superior robustness to alternatives such as Newton's method, as arbitrary initializations are not required.