Li Dazi

   Dazi Li is currently a Professor of automatic control and the Vice Dean of the College of Information Science and Technology, Beijing University of Chemical Technology, Beijing, China. 

   She received the Ph.D. degree in engineering from the Department of Electrical and Electronic Systems, Kyushu University, Fukuoka, Japan, in April, 2004. 

   She has been entitled as the Beijing Teaching Master Award since 2019. She is currently an Associate Editor of ISA Transactions. Her research interests include machine learning and artificial intelligence, industrial safety evaluation & fault diagnosis, advanced process control, complex system modelling and optimization, and fractional calculus system. 

   She is the member of the Fault Diagnosis Professional Committee of the Chinese Association of Automation, the Process Control Professional Committee of the Chinese Association of Automation, and the Intelligent Factory Professional Committee of China Instrument and Control Society.

Study and work experience :

2000.10 - 2004.04  Kyushu University, Japan, Ph.D Degree

1992.09 - 1995.07  Beijing University of Chemical Technology, Master Degree

2004.09 - 2011.11  Beijing University of Chemical Technology, Associate Professor

2011.12 - Present   Beijing University of Chemical Technology, Professor

2014.07 - Present   Beijing University of Chemical Technology, DoctoralSupervisor

   In modern process industries, the dynamic characteristics and complexity of processes are continuously increasing, posing challenges to model-based optimization control methods. In addressing the pressing issues within the field of process control, the applicant has developed a framework for reinforcement learning control that integrates both knowledge and data, leveraging the principles of dynamic programming theory. The primary academic innovations of this work are as follows:

   (1) A high-performance framework of inverse reinforcement learning with reward and policy optimization is proposed, addressing the challenging problem of reward design in reinforcement learning to provide theoretical and methodological support for engineering applications of reinforcement learning.

   (2) A complex process system graph deep reinforcement learning application framework is constructed based on a graph representation model of complex process systems and graph deep learning, achieving automated industrial knowledge extraction and reconstruction in the domain of complex process systems through collaborative driving of both prior knowledge and data.

   (3) A multi-objective decision & optimization method of deep reinforcement learning is proposed under safety constraint conditions, providing a novel solution for the intelligent management of complex process systems, with a focus on green, low-carbon, energy-efficient, and productivity-enhancing objectives.

Courses for graduate students are as follows：

1.《Pattern Recognition-Methods and Application》

Courses for undergraduate students are as follows：

1.《Pattern Recognition and Machine Learning》

2.《Artificial Intelligence & Automation》

3.《Control Engineering Course Design》

4.《Process Control Engineering》

5.《Introduction of Automation Science》