Skip to content

DATA.md

Latest commit

 

History

History
53 lines (30 loc) · 1.58 KB

DATA.md

File metadata and controls

53 lines (30 loc) · 1.58 KB

Instance Preparation

The structure of datasets are organized under ./data as follows

./data/$DATASET_NAME
├── train
    └── instances  # training instances

where each dataset is placed in their own $DATASET_NAME directory

The instances (.mps or .lp files) of each dataset should be prepared and placed in ./data/$DATASET_NAME/train/instances, respectively.

  • For BIP dataset, instances can be downloaded here. One can use instances 0-299 for training.

  • For BPP dataset, instances are generated by running

    python ins_generation.py

    Note: A random seed setting is currently missing; one may add it to address this omission.

  • For SMSP dataset, please convert instances from the steelmillslab set (a copy is available in data/steel) into .mps format by running

    python gen_smsp.py

Solution Collection

The (high-quality) solutions of training instances are collected and stored in ./data/$DATASET_NAME/train/solutions using MILP solver SCIP by

# BPP
python collect_sols.py --rootDir ./data/BPP/train --nWorkers 5 --maxTime 3600


# BIP
python collect_sols.py --rootDir ./data/BIP/train --nWorkers 5 --maxTime 3600


# SMSP
python collect_sols.py --rootDir ./data/SMSP/train --nWorkers 5 --maxTime 3600

Note: The solution files for the BIP and SMSP datasets are already available in their respective directories. Running this step from scratch can take multiple days.