EU Regional School - Laird Seminar
Prof. Dr. Carl D. Laird - Efficient Solution of Large-Scale Nonlinear Programming Problems
Artie McFerrin Department of Chemical Engineering, Texas A&M University
Large-scale nonlinear programming (NLP) has proven to be a fundamental technology for steady state and dynamic optimization of chemical processes. However, as we deal with systems discretized in time and space, include more of the enterprise in a single optimization problem, and seek to rigorously address uncertainty, NLP problems continue to grow increasingly large. This lecture will present the fundamentals of large-scale nonlinear programming, cover the algorithm details of both SQP and interior-point methods, and discuss algorithm specializations necessary for numerical efficiency on large-scale problems. Specific applications will be addressed, including design under uncertainty, and optimization of dynamic systems.