Research Focus
Deep Reinforcement Learning and Safe Decision-Making

A learned control policy is only useful in the field if it never violates transformer ratings, voltage limits, or battery state-of-charge bounds. I develop deep reinforcement learning methods that embed such hard constraints directly into training, together with techniques for reducing very large action spaces, coordinating multiple agents, and transferring policies across sites. We are also beginning to explore what large language models can contribute to user behavior modeling and load forecasting.


Copyright All Rights Reserved Shandong University Address: No. 27 Shanda South Road, Jinan City, Shandong Province, China: 250100
Information desk: (86) - 0531-88395114
On Duty Telephone: (86) - 0531-88364731 Construction and Maintenance: Information Work Office of Shandong University