Welcome to Berkeley Computational Optimization Lab!
Research activities at BCOL are concentrated predominantly on mathematical programming with discrete decision variables, commonly referred to as integer programming. Integer programming is a powerful and versatile modeling and algorithmic approach for optimization problems with discrete choices for the decision variables. These computationally challenging, non-convex optimization problems arise in diverse applications ranging from genetics to financial risk management, from network security to power generation and transmission.
We are in a very exciting period in integer programming research. Recent innovations in optimization theory and algorithms coupled with the advances in computer technology enable us to solve large-scale practical optimization problems that we could not have imagined attacking a decade ago. Researchers in our group develop cutting-edge theories and algorithms that push the limits of solving large-scale integer programs.
Recent research projects of our group include risk-return optimization and utility maximization with binary decisions, conic integer programming, robust optimization, polyhedral cutting planes for mixed-integer programming, superadditive lifting techniques, optimization of logistics networks, polyhedral methods for lot sizing, network flow and design under uncertainty, and survivable network design.
We encourage you to visit BCOL Research Reports page for detailed information on the research projects.
A few researcher positions at doctoral and postdoc levels are available. We are especially interested in applicants with training and/or prior research experience in polyhedral theory, conic optimization, integer programming, and network design. We welcome visiting scholars with similar interests. Please contact Professor Atamturk for further information.
BCOL is equipped with Linux workstations having access to Berkeley DECF Linux computing clusters with more than 200 hosts. The complete suite of IBM CPLEX optimization system and callable libraries, Mosek, SeDuMi, Matlab, and AMPL modeling system, R statistics package are available for the researchers of the lab.
Research projects at BCOL are supported, in part, by grants from the National Science Foundation, the US Department of Defense, the US Department of Energy, the IBM Corporation, and the University of California-Berkeley.