(200d) Optimal Sensitivity Based On IPOPT
AIChE Annual Meeting
2011
2011 Annual Meeting
Computing and Systems Technology Division
Advances In Optimization II
Tuesday, October 18, 2011 - 9:45am to 10:10am
Optimal Sensitivity Based on IPOPT
The sIPOPT library is an open source implementation that provides optimal sensitivity of solutions of nonlinear programming (NLP) problems. It has been adapted to a fast solver based on a barrier NLP method. The program evaluates the sensitivity of the KKT system with respect to model parameters. The code is integrated to the NLP solver IPOPT, and thus we can use the factorized KKT matrix at the solution provided by the solver so that sensitivities to parameters are determined with minimal computational cost. Aside from estimating sensitivities for parametric NLPs, the program provides approximate NLP solutions that can be used for nonlinear model predictive control and state estimation. Moreover, we also provide the elements of a QP-based sensitivity approach through the use of a “fix-relax strategy” to account for active set changes. These are enabled by prefactored KKT matrices and a fix-relax strategy based on Schur complements. In addition, reduced Hessians are obtained at minimal cost and these are particularly effective to approximate covariance matrices in parameter and state estimation problems. The sIPOPT program is demonstrated using the AMPL interface on a simple example. In addition sIPOPT applications to parameter estimation, state estimation and nonlinear model predictive control are highlighted.
Literature cited
Pirnay, H., Lopez-Negrete, R., and Biegler, L.T., Optimal Sensitivity Based on IPOPT, submitted to Math. Prog. Comp., 2011.
Wächter, A. and Biegler, L.T., On the Implementation of a Primal-Dual Interior Point Filter Line Search Algorithm for Large-Scale Nonlinear Programming. Mathematical Programming 106(1), 2006, pp. 25–57.
Fiacco, A.V., Introduction to Sensitivity and Stability Analysis in Nonlinear Programming, volume 165 of Mathematics in Science and Engineering. Academic Press, 1983.
Zavala, V.M., Computational Strategies for the Operation of Large-Scale Chemical Processes. Ph.D. thesis, Carnegie Mellon University, 2008.