Primal Heuristics for Mixed Integer Programs

194 citations recorded by the source corpus.

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

Learning to Run Heuristics in Tree Search (2017, background): “An interesting finding reported in [Berthold, 2006] is that on 97 easy benchmark instances, the LP relaxation finds an optimal solution 59 times, whereas on 26 hard instances it finds an optimal solution only 3 times; for the remaining instances, one of the primal heuristics of the SCIP solver used…”
The SCIP Optimization Suite 3.2 (2016, background): “A description of the existing primal and objective diving heuristics in SCIP can be found in [1, 11, 43].”
Rounding and Propagation Heuristics for Mixed Integer Programming (2011, methodology): “, [2]), propagation heuristics do not solve any LP during search.”
Algorithms and Software for Convex Mixed Integer Nonlinear Programs (2012, methodology): “For MINLP, Cezik and Iyengar [34] demonstrate that if the nonlinear ha l-0 04 23 41 6, v er si on 2 - 15 O ct 2 00 9”
Solving Mixed Integer Programs Using Neural Networks (2020, methodology): “The learned policy significantly outperforms SCIP’s branching heuristic on four datasets with 2-20× better average dual gap (Berthold 2006) on held out instances at large time limits, and has comparable performance on the other rest.”
SCIP: solving constraint integer programs (2009, background): “See Berthold [26] for a detailed discussion of primal heuristics in SCIP .”