Global optimization of bilinear engineering design models
Recently Quesada and Grossmann have proposed a global optimization algorithm for solving NLP problems involving linear fractional and bilinear terms. This model has been motivated by a number of applications in process design. The proposed method relies on the derivation of a convex NLP underestimator problem that is used within a spatial branch and bound search. This paper explores the use of alternative bounding approximations for constructing the underestimator problem. These are applied in the global optimization of problems arising in different engineering areas and for which different relaxations are proposed depending on the mathematical structure of the models. These relaxations include linear and nonlinear underestimator problems. Reformulations that generate additional estimator functions are also employed. Examples from process design, structural design, portfolio investment and layout design are presented.
- OSTI ID:
- 36092
- Report Number(s):
- CONF-9408161--
- Country of Publication:
- United States
- Language:
- English
Similar Records
Solving dynamic control problems via deterministic global optimization
New advances in the GOP algorithm for global optimization