Recursive central rounding for mixed integer programs

Authors:
Joe Naoum-Sawaya
Affiliations:
-
Venue:
Computers and Operations Research
Year:
2014

Citing 9
Cited 0

Decomposition and nondifferentiable optimization with the projective algorithm

Management Science
Interior point algorithms: theory and analysis

Interior point algorithms: theory and analysis
A Nonlinear Analytic Center Cutting Plane Method for a Class of Convex Programming Problems

SIAM Journal on Optimization
The feasibility pump

Mathematical Programming: Series A and B
ZI round, a MIP rounding heuristic

Journal of Heuristics
An interior point cutting plane heuristic for mixed integer programming

Computers and Operations Research
The submodular knapsack polytope

Discrete Optimization
A feasibility pump heuristic for general mixed-integer problems

Discrete Optimization
Improving the feasibility pump

Discrete Optimization

Quantified Score

Hi-index	0.01

Visualization

Abstract

We introduce a new rounding heuristic for mixed integer programs. Starting from a fractional solution, the new approach is based on recursively fixing a subset of the discrete variables while using the analytic center to re-center the remaining ones. The proposed rounding approach can be used independently or integrated with other heuristics. We demonstrate both setups by first using the proposed approach to round the optimal solution of the linear programming relaxation. We then integrate the proposed rounding heuristic with the feasibility pump by replacing the original simple rounding function of the feasibility pump. We conduct computational testing on mixed integer problems from MIPLIB and CORAL and on mixed integer quadratic problems from MIQPLIB. The proposed algorithm is computationally efficient and provides good quality feasible solutions.