Foundation Models for Demand Forecasting via Dual-Strategy Ensembling
This work addresses demand forecasting for supply chain optimization, offering an incremental improvement through ensemble methods to boost generalization in complex environments.
The paper tackled the problem of demand forecasting in supply chains by proposing a dual-strategy ensemble framework that enhances foundation models, resulting in consistent outperformance of baselines and improved accuracy across hierarchical levels in experiments on the M5 benchmark and three sales datasets.
Accurate demand forecasting is critical for supply chain optimization, yet remains difficult in practice due to hierarchical complexity, domain shifts, and evolving external factors. While recent foundation models offer strong potential for time series forecasting, they often suffer from architectural rigidity and limited robustness under distributional change. In this paper, we propose a unified ensemble framework that enhances the performance of foundation models for sales forecasting in real-world supply chains. Our method combines two complementary strategies: (1) Hierarchical Ensemble (HE), which partitions training and inference by semantic levels (e.g., store, category, department) to capture localized patterns; and (2) Architectural Ensemble (AE), which integrates predictions from diverse model backbones to mitigate bias and improve stability. We conduct extensive experiments on the M5 benchmark and three external sales datasets, covering both in-domain and zero-shot forecasting. Results show that our approach consistently outperforms strong baselines, improves accuracy across hierarchical levels, and provides a simple yet effective mechanism for boosting generalization in complex forecasting environments.