The Setting
In dimension , let be a compact convex body of positive Lebesgue volume, and let be compact. The source law is normalized volume on , and the target laws are probability measures supported on . Quadratic transport compares the cost of moving this source by a map with the costs of all couplings having the same marginals.
The Definition
Definition 1. The ambient space is the Euclidean space in (1), with its usual norm and Borel sets: (1)✦
Write for Lebesgue measure on . Write for the source law in (2): (2)✦
A measure on is supported on when it satisfies (3): (3)✦
A probability coupling of measures and on is a probability measure on with marginals and .✦ For measures and on , the set of probability couplings is the set in (4), where are the coordinate projections: (4)✦
The quadratic transport cost of a measure on is the extended nonnegative integral in (5): (5)✦
For , the real representative is the real number specified by when is finite and . The quadratic Wasserstein distance between measures and on is the real-valued quantity in (6): (6)✦
The graph transport plan of a map and a measure on is the pushforward in (7): (7)✦
A quadratic optimal map from to is a measurable map with pushforward whose graph plan minimizes quadratic cost, as in (8): (8)✦
A unique quadratic optimal map from to is a quadratic optimal map for which every coupling of no greater cost is its graph plan, as specified by (9): (9)✦
The distance between maps is the real-valued quantity in (10): (10)✦
What Is Checked
These definitions concern measures and maps on finite-dimensional Euclidean space, with the source measure given by normalized Lebesgue restriction to .