A one-sided constrained martingale transport between two uniform laws

Erhan Bayraktar; Xin Zhang

In the authors' words

We minimize $E [h (Y - X)]$ over martingale couplings of $U ni f [- 1, 1]$ and $U ni f [- 2, 2]$ satisfying $Y \geq X - k$ , where $h \in C^{1} ([- 3, 3])$ has convex derivative. Feasibility holds exactly for $k \geq 1$ . For each such $k$ , we construct a coupling that minimizes all costs in this class. For $1 < k < 3$ , its support consists of two graphs, with $D (x) = x - k$ on $[k - 2, 1]$ . The maps admit an explicit parametrization for $2 \leq k < 3$ and are determined by scalar equations with unique admissible roots for $1 < k < 2$ . We prove optimality by a dual inequality, using an analytic estimate in the latter range. At $k = 1$ the optimizer is $Y = X \pm 1$ with equal probabilities; for $k \geq 3$ it is the ordinary left-curtain coupling. In the unconstrained problem, left-monotonicity identifies the left-curtain coupling, which is optimal for this cost class. Under the constraint, a discrete example shows that the corresponding support condition, even together with every two-source comparison, does not imply optimality.

Main resultLimitation the authors admit

A one-sided constrained martingale transport between two uniform laws

In the authors' words

More on Probability that day

Eigenvalue and Eigenvector Approximation for Random Matrices Using Low-Degree Polynomials

Stopping models closed under pgf composition, and the stability of randomly stopped model extensions

On Fast-Slow Mean-Field Forward-Backward Stochastic Systems

Recurrence and transience of random walks on monotonically changing environments

Shrinking-Tube Concentration for Adaptive Markovian Stochastic Approximation