LOGMIP: a Disjunctive 0-1 Nonlinear Optimizer for Process Systems Models (Full Paper)

ALDO VECCHIETTI; IGNACIO GROSSMANN

COMPUTERS AND CHEMICAL ENGINEERING

Elsevier Science Ltd

Lugar: Grat Britain; Año: 1999 vol. 23 p. 555 - 565

0098-1354

Discrete-continuous non-linear optimization models are frequently used to formulate problems in Process System Engineering.  Major modeling alternatives and  solution algorithms include generalized disjunctive programming and MINLP. Both have advantages and drawbacks  depending on the problem they are dealing with. In this work, we describe the theory  behind LOGMIP, a new computer code for disjunctive programming and MINLP. We discuss a hybrid modeling framework that combines both approaches, allowing binary variables and disjunctions for expressing discrete choices. An extension of the Logic-Based OA algorithm has been implemented to solve the proposed hybrid model. Computational experience is reported on several examples, which are solved using disjunctive, MINLP and hybrid formulations.