From Knothe's Transport to Brenier's Map and a Continuation Method for Optimal Transport

  title={From Knothe's Transport to Brenier's Map and a Continuation Method for Optimal Transport},
  author={Guillaume Carlier and Alfred Galichon and Filippo Santambrogio},
  journal={SIAM J. Math. Anal.},
A simple procedure to map two probability measures in Rd is the so-called Knothe-Rosenblatt rearrangement, which consists in rearranging monotonically the marginal distributions of the last coordinate, and then the conditional distributions, iteratively. We show that this mapping is the limit of solutions to a class of Monge-Kantorovich mass transportation problems with quadratic costs, with the weights of the coordinates asymptotically dominating one another. This enables us to design a… 

Figures from this paper

From Knothe's Rearrangement to Brenier's Optimal Transport Map
It is proved that on the torus (to avoid boundary issues), when all the data are smooth, the evolution is also smooth and is entirely determined by a PDE for the Kantorovich potential (which determines the map) with a subtle initial condition.
Multi-physics optimal transportation and image interpolation
How some physics can be added to the optimal transportation theory, how to construct algorithms to compute solutions to the corresponding optimization problems and how to apply the proposed methods to image interpolation are studied.
Benamou-Brenier and other continuous numerical methods
In this chapter we present some numerical methods to solve optimal transport problems. The most famous method is for sure the one due to J.-D. Benamou and Y. Brenier, which transforms the problem
Stochastic optimal transportation problem and related topics
  • Mathematics, Computer Science
  • 2014
The Knothe-R Rosenblatt process is introduced as a stochastic optimal control analogue of the KnotheRosenblatt rearrangement and it is shown that it can be approximated by a minimizer of a class of Stochastic optimal transportation problems with a small parameter.
A Newton algorithm for semi-discrete optimal transport
The purpose of this article is to bridge the gap between theory and practice by introducing a damped Newton's algorithm which is experimentally efficient and by proving the global convergence of this algorithm with optimal rates.
Causal Transport in Discrete Time and Applications
A dynamic programming principle is established that links the causal transport problem to the transport problem for general costs recently considered by Gozlan et al. and gives conditions under which the celebrated Knothe-Rosenblatt rearrangement can be viewed as a causal analogue to the Brenier's map.
An algorithm for optimal transport between a simplex soup and a point cloud
The convergence with linear speed of a damped Newton's algorithm is proved to solve the optimal transport problem as the resolution of a non-linear system where one wants to prescribe the quantity of mass in each cell of the so-called Laguerre diagram.
A Hierarchical Approach to Optimal Transport
An extension of the auction algorithm is presented that exploits the regularity of the otherwise arbitrary cost function and takes into account a sparse subset of possible assignment pairs while still guaranteeing global optimality of the solution.
Optimal transport and barycenters for dendritic measures
A variant of the Wasserstein distance on the space of probability measures, specially designed to deal with measures whose support has a dendritic, or treelike structure with a particular direction of orientation, is introduced.
Particle flow inspired by Knothe-Rosenblatt transport for nonlinear filters
We derive a new algorithm for particle flow corresponding to Bayes’ rule that was inspired by Knothe- Rosenblatt transport, which is well known in transport theory. We emphasize that our flow is not


Topics in Optimal Transportation
Introduction The Kantorovich duality Geometry of optimal transportation Brenier's polar factorization theorem The Monge-Ampere equation Displacement interpolation and displacement convexity Geometric
Existence and uniqueness of monotone measure-preserving maps
Introduction. Given a pair of Borel probability measures # and v on Rd, it is natural to ask whether v can be obtained from # by redistributing its mass in a canonical way. In the case of the line d
Mass transportation problems
Modifications of the Monge-Kantorovich Problems: Transportation Problems with Relaxed or Additional Constraints.- Application of Kantorovich-Type Metrics to Various Probabilistic-Type Limit
An elementary proof of the polar factorization of vector-valued functions
We present an elementary proof of an important result of Y. Brenier [Br1, Br2], namely, that vector fields in ℝd satisfying a nondegeneracy condition admit the polar factorization (*) u(x)=▽ψ(s(x)),
Polar Factorization and Monotone Rearrangement of Vector-Valued Functions
Given a probability space (X, p) and a bounded domain R in Rd equipped with the Lebesgue measure 1 . I (normalized so that 10 I = I ), it is shown (under additional technical assumptions on X and Q)
A computational fluid mechanics solution to the Monge-Kantorovich mass transfer problem
Summary. The $L^2$ Monge-Kantorovich mass transfer problem [31] is reset in a fluid mechanics framework and numerically solved by an augmented Lagrangian method.
Contributions to the theory of convex bodies.
Remarks on a Multivariate Transformation
In connection with the study of interpersonal verbal communication, the Washington Public Opinion Laboratory designed an experimental procedure which yielded data on the temporal diffusion of thirty-three different messages in a population of 184 individuals.
Topics in Optimal Transportation, Providence
  • American Mathematical Society,
  • 2003
Optimal transport: Old and New, lecture notes, Ecole d'´ eté de probabilités de Saint-Flour, to appear
  • Optimal transport: Old and New, lecture notes, Ecole d'´ eté de probabilités de Saint-Flour, to appear
  • 2008