EURO 2024 Copenhagen
Abstract Submission

EURO-Online login

1621. Bilinear Branch and Check Algorithm

Invited abstract in session WC-4: Applications of Mixed-Integer and Nonconvex Optimization 1, stream MINLP.

Wednesday, 12:30-14:00
Room: 1001 (building: 202)

Authors (first author is the speaker)

1. Ponpot Jartnillaphand
School of Electrical Engineering, Computing and Mathematical Sciences, Curtin University
2. Hoa Bui
Curtin Centre for Optimisation and Decision Science, Curtin University

Abstract

In this presentation, we introduce the bilinear branch and check algorithm, an exact method for solving nonlinear optimization problems. This approach extends the classical branch and check algorithm by integrating bilinear constraints into the search process. The enhancement of the bilinear constraint refines the problem formulation, expediting the convergence of the search. The algorithm decomposes problems into manageable subproblems, utilizing the branch and cuts to assess feasibility and generate feasibility cuts (no-good cuts), while also enhancing master problem solutions. A key innovation lies in the algorithm's methodology for improving master problem solutions and constructing high-quality no-good cuts among decision variables. This allows for further decomposition of the master problem, enhancing its effectiveness. Our numerical results demonstrate promising outcomes and exhibit greater efficiency compared to standard branch and cut methods like IBM CPLEX and the classical branch and check.

Keywords

Status: accepted


Back to the list of papers