Martingales & transport · Stochastic control / Open question
Extremal structure and certified bounds in multidimensional martingale transport
Turns observable option-price constraints into rigorous bounds for multi-asset exotic derivatives without choosing a single calibrated stochastic model.
Seed catalog · Showing the bundled record.
Problem statement
For vector-valued asset prices at finitely many dates, characterize the geometry of martingale laws that maximize or minimize the expected value of a specified multi-asset payoff subject to marginal constraints. Develop convergent computational constructions of these extremal laws and matching sub- or superhedges, with certified errors for the resulting model-independent derivative bounds as discretization is refined.
Assumptions and scope
- Specify whether the data fix full joint asset marginals at each date or only each asset's individual marginals; these define different feasible sets.
- Assume a nonempty class of integrable martingale laws, and fix payoff continuity or semicontinuity, growth and moment conditions.
- Use at least two assets; specify the time grid, asset dimension and payoff family when seeking structural or complexity theorems.
- State admissible static option positions and predictable dynamic strategies in the dual hedging problem; do not assume dual attainment without its hypotheses.
What would settle it?
New structural characterizations for a specified multidimensional payoff/marginal class and an algorithm with matching primal-dual certificates and a proved continuum error bound; explain dependence on dimension and number of dates.
Selected for a theoretical or computational target. A successful calibration or numerical benchmark alone does not establish the stated mathematical guarantee.
Status and evidence
Existence, duality and extremal-support results are known in substantial regimes. Ghoussoub, Kim and Lim establish higher-dimensional structural results under specific hypotheses. Che, Lim and Sun's February 2026 preprint reports multi-period, multi-asset dual attainment under regularity and irreducibility assumptions, with discrete computations. The remaining target is broader extremal geometry and certified scalable approximation; it does not label general duality or the cited attainment theorem as unsolved.
Preliminary source check; not an expert-verified open status or an independent audit of proposed proofs. The compilation date is not a literature-review cutoff.
Sources and further reading
- Ghoussoub, Kim & Lim: Structure of optimal martingale transport plans in general dimensions ↗partial progress · source checked 2026-11-09
Scholarly abstract: increment-modulus costs, dual attainment and extreme-point support results.
The structural conclusions depend on dimension, marginal order and regularity.
- Che, Lim & Sun: Dual Attainment in Multi-Period Multi-Asset Martingale Optimal Transport and Its Computation (2026 preprint) ↗partial progress · source checked 2026-11-09
arXiv abstract: regularity/irreducibility conditions and discrete PDLP experiments.
Recent preprint claim; the abstract was checked, not the full proof or continuum numerical error.