From Shared Demand Patterns to Local Uncertainty: Probabilistic Load Forecasting by Mixing Compact Adaptations

2026-10-06

AI Summary

This study addresses the challenge in customer-level load forecasting where single models struggle to capture heterogeneous patterns while independent modeling incurs prohibitive costs. To this end, we propose a scalable probabilistic forecasting framework that leverages a shared model architecture to learn global common features and employs a mixture of low-rank compact adapter components to achieve local personalized adaptation, effectively balancing knowledge sharing with heterogeneity modeling. Experiments conducted on 590 load profiles from the SMART-DS dataset demonstrate that the proposed framework significantly improves both deterministic and probabilistic forecasting accuracy while maintaining minimal parameter storage and inference overhead.

Abstract

Probabilistic load forecasting has been widely studied for power-system operation and planning, but customer- and transformer-level forecasting introduces a distinct scalability challenge. At these levels, load uncertainty is strongly affected by customer behavior, weather, and mixed load composition, making it difficult for a single shared model to capture heterogeneous patterns. Using separate probabilistic models can improve local accuracy, but becomes costly to train, store, update, and validate at scale. To address this challenge, we develop a scalable customer-aware forecasting framework that learns common demand behavior through a shared model while adapting only a compact subset of parameters. Rather than using an independent model for each load or assigning each load to a specialized model, the proposed design learns a small bank of low-dimensional adaptation components and allows each load to combine them according to its forecasting characteristics. This preserves shared knowledge across customers while providing sufficient flexibility for heterogeneous and mixed load compositions. Experiments on 590 load profiles from the SMART-DS dataset show consistent improvements in deterministic accuracy and probabilistic quality over statistical, neural-network, Transformer-based, and pretrained time-series baselines, while retaining low storage and inference costs.

Problem

• Probabilistic load forecasting
• Scalability
• Local uncertainty
• Heterogeneous patterns

Innovation

• Probabilistic Load Forecasting
• Compact Adaptations
• Scalable Framework
• Shared Model
• Low-dimensional Components