Federated learning method handles varied decision tasks across clients
FedRSPO+: A Heterogeneity-aware Algorithm for Decision-focused Federated Learning
Machine Learning
Summary
Training models to make good decisions can be tricky when data is spread across different places with different rules. The authors study this problem where each client’s decision goals and limitations can vary a lot. They introduce FedRSPO+, a new way to smooth out these differences so the training stays stable and effective. Their experiments show this method helps multiple clients learn jointly to make better decisions, even when their situations differ.
What this means in practice
- •For energy management teams: Improve collaborative energy pricing models across different regions with varying constraints and objectives using stable federated decision-focused learning.
- •For logistics planners: Coordinate route optimization across distributed locations with differing needs by smoothing decision updates in federated learning.
Authors
Konstantinos Ziliaskopoulos, Alexander Vinel, Jiaqi Wang
Abstract
Decision-focused learning (DFL) trains predictive models for downstream optimization, but existing methods largely assume centralized data. In cross-silo settings, federated learning offers a natural alternative, yet standard federated methods optimize prediction over decision quality and do not address heterogeneity in downstream objectives or feasible sets. This heterogeneity is especially challenging for DFL because small perturbations in polyhedral problems can cause discontinuous changes in optimal decisions, destabilizing client updates and aggregation. We propose FedRSPO+, a heterogeneity-aware framework for decision-focused federated learning, built on RSPO+, a regularized predict-then-optimize surrogate that smooths the decision map through projection. We show that RSPO+ upper bounds decision error and regret for the regularized decision and, under exact regularization and consistent LP solution selection, for the original LP decision. We further derive cross-client heterogeneity bounds that depend on both objective and feasible-set heterogeneity, vanish at homogeneity, and require no strong convexity. FedRSPO+ uses an annealed, modular training procedure compatible with standard federated personalization and aggregation methods. Experiments on synthetic knapsack, shortest-path, and real-world energy pricing tasks compare against prediction-only federated learning and DFL baselines under varying heterogeneity and communication budgets. Results suggest that smoothing is a useful ingredient for stable collaborative decision learning and provide a heterogeneity-aware foundation for federated DFL.