Diffusion-Based Forecasting for Uncertainty-Aware Model Predictive Control

Stelios Zarifis, Ioannis Kordonis, Petros Maragos, [0.5em] Robotics Institute, Athena Research Center, Athens, Greece · 2025

Read Diffusion-Based Forecasting for Uncertainty-Aware Model Predictive Control by Stelios Zarifis, Ioannis Kordonis, Petros Maragos, [0.5em] Robotics Institute, Athena Research Center, Athens, Greece (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2503-15095.txt.

Description

Stelios Zarifis, Ioannis Kordonis, and Petros Maragos blend image-style generators into control math in a 2025 paper. Model predictive control steers a system by looking ahead, but the future is hazy when parts of the system are random and half-hidden. Their frame feeds the controller a range of likely futures, not one guess.

A diffusion model, borrowed from time-series forecasting, draws those futures as samples. The controller then picks moves that fare well across the spread. Tested on battery trading in New York's day-ahead power market, the method beat both classic forecast-driven control and learning-based rivals.

Catalog record

CREATORStelios Zarifis, Ioannis Kordonis, Petros Maragos, [0.5em] Robotics Institute, Athena Research Center, Athens, Greece
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE25 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2503-15095

Browse the full library · About the Tower of Babel