Grand-Canonical Optimal Transport

Simone Di Marino,Mathieu Lewin,Luca Nenna
2024-09-16
Abstract:We study a generalization of the multi-marginal optimal transport problem, which has no fixed number of marginals $N$ and is inspired of statistical mechanics. It consists in optimizing a linear combination of the costs for all the possible $N$'s, while fixing a certain linear combination of the corresponding marginals.
Optimization and Control,Statistical Mechanics,Mathematical Physics
What problem does this paper attempt to address?
This paper attempts to address the problem of studying a more general multi-marginal optimal transport problem within the framework of Grand-Canonical Optimal Transport (GC-OT). Specifically, the paper explores an optimal transport problem with a non-fixed number of marginals, a model referred to as the "Grand-Canonical Ensemble" in statistical physics. The main contributions of the paper are as follows: 1. **Mathematical Properties of the Grand-Canonical Optimal Transport Problem**: - Investigated the fundamental mathematical properties of the GC-OT problem, including existence and uniqueness. - Proved the existence of an optimal solution under appropriate cost assumptions. 2. **Case of Symmetric Costs**: - Paid special attention to the case of pairwise costs, where all costs can be represented by the cost between two agents. - Discussed specific examples such as Coulomb interactions in detail. 3. **Grand-Canonical Ensemble Characteristics of Subsystems**: - Explained why it is natural to consider the variation in the number of marginals even when starting from a system with a fixed number of agents. - Discussed how to apply the concept of the Grand-Canonical Ensemble in subsystems. 4. **Stability and Hyper-Stability Conditions**: - Proposed definitions of stability and hyper-stability to ensure the well-definedness of the problem. - Proved the existence of an optimal solution under hyper-stability conditions. 5. **Duality Theory**: - Studied the duality theory of the GC-OT problem, including the existence of dual potentials. 6. **Study of the One-Dimensional Case**: - Conducted a detailed analysis of the one-dimensional case, verifying some predictions about the shape of the optimal transport plan. 7. **Entropy Regularization**: - The final section discussed the entropy-regularized version of the GC-OT problem. In summary, this paper aims to address some limitations of traditional multi-marginal optimal transport problems by introducing the Grand-Canonical Optimal Transport framework and exploring its applications in different scenarios.