On the 18th of September 2017, the KIOS Research and Innovation Center of Excellence (CoE) launched the KIOS Distinguished Lecture Series. The aim of the KIOS Distinguished Lecture Series is to host world-class researchers at the University of Cyprus to deliver specialized lectures in various research, technology, and innovation fields. The goal is to provide the scientific communities in Cyprus with invaluable insights into globally leading research and innovation results. The Series also aims to enrich interaction with leading figures in relevant science and technology fields. The launch of the series was marked by the world-renowned Professor Dimitri Bertsekas from the Massachusetts Institute of Technology, USA.
Professor Bertsekas has been teaching at the Electrical Engineering and Computer Science Department of the Massachusetts Institute of Technology (MIT) since 1979, where he is currently the McAfee Professor of Engineering. Professor Dimitri Bertsekas was awarded with several internationally recognized and prestigious awards and in 2001 was elected to the United States National Academy of Engineering for “pioneering contributions to fundamental research, practice and education of optimization/control theory, and especially its application to data communication networks”. He has written numerous research papers, and sixteen books and research monographs, several of which are used as textbooks in MIT classes.
The Approximate Dynamic Programming (DP)
The topic of Professor Bertsekas’ lecture was on “Approximate Dynamic Programming”, an area of significant relevance for a variety of scientific fields. The Dynamic Programming (DP) algorithm for stochastic sequential decision problems is utilized in a broad variety of applications, such as control, robotics, planning, operations research, economics, artificial intelligence, and others. This is a technique for modelling and solving problems of decision making under uncertainty, aiming to compute a policy prescribing how to act optimally in the face of uncertainty. Very often, because of overwhelming computational requirements, it is practical to settle for suboptimal control schemes that strike a reasonable balance between practical implementation and adequate performance. Focusing on this fact, Prof. Bertsekas reviewed several approaches for suboptimal control, addressing approximations both in value and in policy space. The seminar concluded with the description of some of the currently popular approximate DP schemes, and reviewed their range of applications.
Prof. Bertseka’s lecture was extremely well received, attracting an audience of approximately 150 attendees, including several academics and researchers from the University of Cyprus, as well as other Universities in the country. The lecture was also attended by a number of experts from industry, including collaborators of the KIOS CoE from various government organizations and high-tech companies.
The full videos of the lecture series can be found here.