Threshold-Based Algorithms for an Online Rolling Horizon Framework Under Uncertainty -- With an Application to Energy Management

TL;DR

This paper introduces an online scheduling algorithm for a rolling horizon framework that leverages short-term forecasts and observations to effectively manage uncertainty, significantly reducing energy costs in microgrid applications.

Contribution

It presents a novel online algorithm based on combinatorial online optimization and robust optimization principles for dynamic energy management under uncertainty.

Findings

Reduces microgrid electricity costs by over 85% compared to classical methods.

Achieves more than 50% cost reduction over offline dynamic frameworks.

Provides detailed analysis of algorithm performance under various forecast errors.

Abstract

Decision problems encountered in practice often possess a highly dynamic and uncertain nature. In particular fast changing forecasts for parameters (e.g., photovoltaic generation forecasts in the context of energy management) pose large challenges for the classical rolling horizon framework. Within this work, we propose an online scheduling algorithm for a rolling horizon framework, which directly uses short-term forecasts and observations of the uncertainty. The online scheduling algorithm is based on insights and results from combinatorial online optimization problems and makes use of key properties of robust optimization. Applied within a robust energy management approach, we show that the online scheduling algorithm is able to reduce the total electricity costs within a local microgrid by more than 85% compared to a classical rolling horizon framework and by more than 50% compared to a tailor-made dynamic, yet still offline rolling horizon framework. A detailed analysis provides insights into the working of the online scheduling algorithm under different underlying forecast error distributions.