Skip to content

thuiar/MILPGen

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

3 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

MILPGen

This is the official code implementation of paper "Learning to Generate Scalable MILP instances."

Environment Setup

conda env create -f scripts/environment.yml
conda activate MILPGen

Code

There are two stages of the code: the first one is instance clustering, and the second one is instance generating.

Instance Clustering

Create directory and generating benchmark problems:

mkdir bipartite_graph
mkdir tmp_result
python 01.py --problem_type=IS --number=20 --output_dir=bipartite_graph/
python 01.py --problem_type=MVC --number=20 --output_dir=bipartite_graph/
python 01.py --problem_type=CAT --number=20 --output_dir=bipartite_graph/

Train VGAE to get problem embeddings and clustering instances:

python 02.py --input_dir=bipartite_graph/ --epoch=10 --output_file=tmp/02

python 03.py --class_num=3 --input_file=tmp/02 --output_file=tmp/03

The code solve_gurobi.py and solve_scip.py provides an entry point to solve dumped bipartite representation of MILP instances.

Instance Generating

Instance Generating is intend for single category of problems.

We refer to scripts under scripts/ directory.

(Scripts are under construction)

Citations

If this work is helpful, or you want to use the codes and results in this repo, please cite the following papers:

(To be appeared)

About

No description, website, or topics provided.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages