Branching and bounds tightening techniques for non-convex MINLP

被引:363
作者
Belotti, Pietro [1 ]
Lee, Jon [2 ]
Liberti, Leo [4 ]
Margot, Francois [3 ]
Waechter, Andreas [2 ]
机构
[1] Lehigh Univ, Dept Ind & Syst Engn, Bethlehem, PA 18015 USA
[2] IBM Corp, TJ Watson Res Ctr, Yorktown Hts, NY USA
[3] Carnegie Mellon Univ, Tepper Sch Business, Pittsburgh, PA 15213 USA
[4] Ecole Polytech, LIX, F-91128 Palaiseau, France
基金
美国国家科学基金会;
关键词
mixed-integer non-linear programming; Couenne; branching rules; bounds tightening; GLOBAL OPTIMIZATION METHOD; DIFFERENTIABLE CONSTRAINED NLPS; INTEGER NONLINEAR PROGRAMS; MIXED-INTEGER; ALPHA-BB; INTERVAL-ANALYSIS; CUT ALGORITHM; DESIGN; IMPLEMENTATION; MINIMIZATION;
D O I
10.1080/10556780903087124
中图分类号
TP31 [计算机软件];
学科分类号
081202 ; 0835 ;
摘要
Many industrial problems can be naturally formulated using mixed integer non-linear programming (MINLP) models and can be solved by spatial BranchBound (sBB) techniques. We study the impact of two important parts of sBB methods: bounds tightening (BT) and branching strategies. We extend a branching technique originally developed for MILP, reliability branching, to the MINLP case. Motivated by the demand for open-source solvers for real-world MINLP problems, we have developed an sBB software package named couenne (Convex Over- and Under-ENvelopes for Non-linear Estimation) and used it for extensive tests on several combinations of BT and branching techniques on a set of publicly available and real-world MINLP instances. We also compare the performance of couenne with a state-of-the-art MINLP solver.
引用
收藏
页码:597 / 634
页数:38
相关论文
共 90 条
[1]   Branching rules revisited [J].
Achterberg, T ;
Koch, T ;
Martin, A .
OPERATIONS RESEARCH LETTERS, 2005, 33 (01) :42-54
[2]  
Adjiman CS, 1997, COMPUT CHEM ENG, V21, pS445
[3]   A global optimization method, alpha BB, for process design [J].
Adjiman, CS ;
Androulakis, IP ;
Maranas, CD ;
Floudas, CA .
COMPUTERS & CHEMICAL ENGINEERING, 1996, 20 :S419-S424
[4]   A global optimization method, αBB, for general twice-differentiable constrained NLPs -: I.: Theoretical advances [J].
Adjiman, CS ;
Dallwig, S ;
Floudas, CA ;
Neumaier, A .
COMPUTERS & CHEMICAL ENGINEERING, 1998, 22 (09) :1137-1158
[5]   Rigorous convex underestimators for general twice-differentiable problems [J].
Adjiman, CS ;
Floudas, CA .
JOURNAL OF GLOBAL OPTIMIZATION, 1996, 9 (01) :23-40
[6]   A global optimization method, αBB, for general twice-differentiable constrained NLPs -: II.: Implementation and computational results [J].
Adjiman, CS ;
Androulakis, IP ;
Floudas, CA .
COMPUTERS & CHEMICAL ENGINEERING, 1998, 22 (09) :1159-1179
[7]  
ADJIMAN CS, 1998, HDB COMBINATORIAL OP, V1, P1
[8]   alpha BB: A global optimization method for general constrained nonconvex problems [J].
Androulakis, IP ;
Maranas, CD ;
Floudas, CA .
JOURNAL OF GLOBAL OPTIMIZATION, 1995, 7 (04) :337-363
[9]  
[Anonymous], 1996, Global Optimization. Deterministic Approaches
[10]  
[Anonymous], 1992, Global optimization using interval analysis