Cornell University

This is a past event. Its details are archived for historical purposes.

The contact information may no longer be valid.

Please visit our current events listings to look for similar events by title, location, or venue.

CAM Colloquium: Fengqi You (Chemical and Biomolecular Engineering, Cornell) - Projection-based Reformulation and Decomposition Algorithm for Mixed-Integer Bilevel Linear Programs

Friday, December 2, 2016 at 3:30pm

Frank H. T. Rhodes Hall, 655

Mixed-Integer Bilevel Linear Program (MIBLP) is a class of most challenging optimization problems that has a bilevel optimization structure and includes integer variables in both upper and lower problems. Despite the growing applications of MIBLPs in various areas, there is no efficient solution algorithm for large-scale MIBLPs; existing MIBLP algorithms are either subject to simplifying assumptions on the integrity of parameters/variables or restrictions on the presence of upper-level connecting constraints. The complexity of bilevel optimization lies in the property that constraint region of the upper-level problem is partially determined by the solutions to a lower-level optimization problem. MIBLP problems are further complicated because 1) the bilevel feasible region can be nonconvex and disconnected; 2) removing the integrality constraints does not necessarily provide a valid relaxation of the original MIBLP problem; 3) lower-level optimal solutions are not always feasible to the original MIBLP when upper-level connecting constraints are present. All these challenges make MIBLPs much more difficult to solve than single-level mixed-integer linear programs, which are known as NP-hard problems. In this talk, I will present recent theoretical, algorithmic and computational results on global optimization of large-scale MIBLPs. After discussing theory and proprieties of MIBLPs, I’ll introduce a novel MIBLP algorithm that has the least restrictions on problem structure and outperforms existing ones by at least several orders of magnitude in terms of computational efficiency.

 

Biography:
Fengqi You is the Roxanne E. and Michael J. Zak Professor in the Smith School of Chemical and Biomolecular Engineering at Cornell University. Prior to his appointment at Cornell in 2016, he served on the faculty of Northwestern University for five years and was an Argonne Scholar at Argonne National Laboratory from 2009 to 2011. He has published over 100 peer-reviewed journal articles, some of which have been editorially highlighted in Nature, featured on journal covers (e.g. Energy & Environmental Science and ACS Sustainable Chemistry & Engineering), and covered by major media outlets (e.g. The New York Times, BBC, BusinessWeek, and National Geographic). He has received several competitive awards, including the 2011 W. David Smith, Jr. Publication Award from American Institute of Chemical Engineers, the 2013 Northwestern-Argonne Early Career Investigator Award, and a CAREER Award from the National Science Foundation in 2016, as well as a number of best paper awards and “most cited” article awards. He is currently an Associate Editor for Computers & Chemical Engineering. Fengqi earned a B.Eng. from Tsinghua University and received his Ph.D. from Carnegie Mellon University, both in chemical engineering. His research focuses on the development of novel computational models, optimization algorithms, and systems analysis tools for process manufacturing, energy systems and sustainability. More information about his research group can be found from the website: <http://you.cbe.cornell.edu>.

Subscribe
Google Calendar iCal Outlook
Event Type

Seminar

Departments

Cornell Engineering, Mathematics, Center for Applied Mathematics

Tags

engineering, math, center for applied math, cam, College of Engineering, mathematics, center for applied mathematics

Website

http://www.cam.cornell.edu

Speaker

Fengqi You

Speaker Affiliation

Roxanne E. and Michael J. Zak Professor, School of Chemical and Biomolecular Engineering, Cornell University

Recent Activity