TITLE: Conic Optimization: the Past, the Present, and the Future
SPEAKER: Dr Imre Pólik, Postdoctoral Research Fellow
School of Computational Engineering and Science,
McMaster University, Canada
DATE / TIME: Thursday, April 17, 2008 / 2:30 - 3:30 p.m.
LOCATION: Room 451 Mohler Lab, 200 W. Packer Avenue
ABSTRACT: Conic optimization faces a major challenge these days. Despite being well-established in the academic world it is yet to make its breakthrough in industrial environments. Several applications exist, but they usually remain within the walls of universities. Furthermore, integration of conic optimization in modeling languages is scarce at best. The goal of this talk is to identify the causes and to sample some ideas for the future. I will show several motivating examples of conic (mainly second-order and semi definite) optimization, and then present the state-of-the-art about algorithms. A discussion of the strengths and weaknesses of available software implementations will identify the key areas which are essential to the widespread use of this framework. I will then present current research directions and future challenges. Elements of the theory will be kept to a minimum.
BIOGRAPHY: Dr Imre Pólik is a postdoctoral research fellow at the School of Computational Engineering and Science, McMaster University, Canada. He earned his PhD at McMaster University in 2007 under the supervision of Dr Tamás Terlaky. His primary research interest is in high performance computing and conic optimization. Recently, he has been increasingly more interested in practical extensions of the theory and efficient implementations of the algorithms. Following the death of Jos Sturm, Dr Pólik and his colleagues took over and continued the development of the celebrated SeDuMi package, a software tool for conic optimization, which he plans to continue in the coming years. Dr Pólik received the 2006 Best Student Paper Award (PhD students) of MITACS, and was a member of Team OptLab winning the 2007 CORS/VISUAL8 simulation competition.
