CENTRALIZED/DECENTRALIZED is an interface designed to solve Generalized Disjunctive Program (GDP) model for the Optimal Location of Centralized and Decentralized Manufacture Facilities based on their maximum capacity and the given coordinates of the suppliers and consumer markets, while minimizing investment, transportation and purchases of raw material costs.
The model involves bilinear terms for the transportation costs, and nonlinear convex constraints for calculating Euclidean distances. A proposed global bilevel decomposition algorithm based on Global Logic-Based Outer Approximation (GLBOA) by Trespalacios and Grossmann (2015) that involves the successive solution of master problems and reduced size subproblems. A MILP master problem that predicts lower bounds is proposed to approximate the corresponding nonconvex MINLP problem. The subproblem corresponds to a reduced size MINLP problem that is solved to global optimality.
The proposed method has been applied to small and large-scale problems. The results show that the algorithm is much more effective at finding optimal or near-optimal solutions than general-purpose global optimization solvers.
CENTRALIZED/DECENTRALIZED is based on the papers by Cristiana L. Lara, Ignacio E. Grossmann, "Global Optimization for a Continuous Location-Allocation Model for Centralized and Distributed Manufacturing", Computer Aided Chemical Engineering. Volume 38, 2016, Pages 1009-1014 , and by Cristiana L. Lara, Francisco Trespalacios & Ignacio E. Grossmann "Global Optimization Algorithm for Capacitated Multi-facility Continuous Location-allocation Problems", Journal of Global Optimization volume 71, pages 871–889 (2018).
CENTRALIZED/DECENTRALIZED has been developed by Rosanna Franco under the supervision of Ignacio E. Grossmann.