Federated Learning for Discrete Optimal Transport with Large Population under Incomplete Information
This work addresses resource allocation problems in large-scale systems, such as logistics or healthcare, but appears incremental as it builds on existing federated learning and optimal transport methods.
The paper tackles the challenge of scaling optimal transport for large, heterogeneous populations by introducing a discrete framework and addressing both known and unknown type distributions, with a federated learning approach for privacy-preserving computation in the unknown case, though no concrete performance numbers are provided.
Optimal transport is a powerful framework for the efficient allocation of resources between sources and targets. However, traditional models often struggle to scale effectively in the presence of large and heterogeneous populations. In this work, we introduce a discrete optimal transport framework designed to handle large-scale, heterogeneous target populations, characterized by type distributions. We address two scenarios: one where the type distribution of targets is known, and one where it is unknown. For the known distribution, we propose a fully distributed algorithm to achieve optimal resource allocation. In the case of unknown distribution, we develop a federated learning-based approach that enables efficient computation of the optimal transport scheme while preserving privacy. Case studies are provided to evaluate the performance of our learning algorithm.