Project Details

Layman's description

Solving large-scale real-time assets scheduling problems in renewable power systems is computationally challenging due to binary nature of some decisions, diverse types of restrictions and the time limit. This issue is getting worse as more renewable energy sources are integrated into power systems, since hedging against their uncertainty necessitates quicker solution methods to make faster sequential operation decisions at rates unreachable by conventional algorithms. In practical settings, real-time scheduling problems are repeatedly solved through mixed-integer optimization, sometimes multiple times per day with only minor changes in input data. As a result, there is a great potential in reducing complexity and boosting the solution process by learning from previous schedules. The goal of this research is to explore how we may use machine learning to improve the computational performance with a performance guarantee.
Effective start/end date01/03/202328/02/2026


Explore the research topics touched on by this project. These labels are generated based on the underlying awards/grants. Together they form a unique fingerprint.