Integration of polyhedral outer approximation algorithms with MIP solvers through callbacks and lazy constraints

A4 Konferenspublikationer


Interna författare/redaktörer


Publikationens författare: Andreas Lundell, Jan Kronqvist
Redaktörer: Michael T. M. Emmerich, André H. Deutz, Sander C. Hille, Yaroslav D. Sergeyev
Förlagsort: Melville, NY
Publiceringsår: 2019
Tidskrift: AIP Conference Proceedings
Förläggare: American Institute of Physics
Moderpublikationens namn: Proceedings LeGO : 14th International Global Optimization Workshop : Leiden, the Netherlands, 18-21 September 2018
Tidskriftsakronym: AIP CONF PROC
Seriens namn: AIP conference proceedings
Volym: 2070
Antal sidor: 4
ISBN: 978-0-7354-1798-4
ISSN: 0094-243X


Abstrakt

In this paper, it is explained how algorithms for convex mixed-integer nonlinear programming (MINLP) based on polyhedral outer approximation (P0A) can be integrated with mixed-integer programming (MIP) solvers through callbacks and lazy constraints. Through this integration, a new approach utilizing a single branching tree is obtained which reduces the overhead required when rebuilding the branching tree in the MIP solver due to the continuous addition of linear constraints approximating the nonlinear feasible region of the MINLP problem. The result is an efficient strategy for implementing a POA utilized by the Supporting llyperplane Optimization Toolkit (SHOT) solver.

Senast uppdaterad 2020-22-02 vid 05:37