A relaxed reduced space SQP strategy for dynamic optimization problems.
Recently, strategies have been developed to solve dynamic simulation and optimization problems in a simultaneous manner by applying orthogonal collocation on finite elements and solving the nonlinear program (NLP) with a reduced space successive quadratic programming (SQP) approach. We develop a relaxed simultaneous approach that leads to faster performance. The method operates in the reduced space of the control variables and solves the collocation equations inexactly at each SQP iteration. Unlike previous simultaneous formulations, it is able to consider the state variables one element at a time. Also, this approach is compared on two process examples to the reduced gradient, feasible path approach outlined in Logsdon and Biegler. Nonlinear programs with up to 5500 variables are solved with only 40% of the effort. Finally, a theoretical analysis of this approach is provided.
- Research Organization:
- Argonne National Lab. (ANL), Argonne, IL (United States)
- Sponsoring Organization:
- ER
- OSTI ID:
- 937624
- Report Number(s):
- ANL/MCS/JA-2840; CCENDW; TRN: US200905%%351
- Journal Information:
- Comput. Chem. Eng., Vol. 17, Issue 4 ; 1993; ISSN 0098-1354
- Country of Publication:
- United States
- Language:
- ENGLISH
Similar Records
Direct SQP-methods for solving optimal control problems with delays
ALESQP: An Augmented Lagrangian Equality-Constrained SQP Method for Optimization with General Constraints