Forecasting, language, contextual signals, and decision support

Forecasting and contextual information in the loop

Forecasting methods that incorporate contextual information and improve downstream decision-making and control in energy systems.

This project direction asks how forecasts should be corrected or refined when they directly affect control and operational performance. It includes language-informed forecast updates, retrieval-based contextual reasoning, and benchmark design for text-informed electricity-price forecasting.

Research Focus

Related Publications

View all publications →
[P3]Conference Paper

LiFT-MPC: Language-in-the-Loop Feedback Tuning of Cost Previews for MPC

X. Yi, I. Lestas

IEEE Conference on Decision and Control (CDC), 2026 (accepted).

[W1]Working Paper

Economic MPC for Battery Energy Storage Systems with Language-Informed Electricity Price Forecast Refinement

X. Yi, I. Lestas

work in preparation.

[W2]Working Paper

EPNews: A Multi-Market Benchmark for Text-Informed Electricity Price Forecasting

X. Yi, I. Lestas

work in preparation.

[W3]Working Paper

SPARK: A Scenario-Based Price Correction Agent with Retrieval Knowledge for Electricity Price Forecast Correction

X. Yi, I. Lestas

work in preparation.