This implementation was used to conduct the experiments in our paper Deep Learning Assisted Heuristic Tree Search for the Container Pre-marshalling Problem (the preprint of the paper can be found here). Additionally to the code, this repository also contains the validation and test instances (and their optimal solutions) used in the paper.
We evaluated DLTS using Python 3.5 and
- keras 1.1.0
- theano 0.8.2
- h5py 2.7.1
To solve the 5x5_cv1 validation instances using the provided branching and bounding networks use the following command:
python deep_learn.py -t reference_solutions/5x5_cv1_validation -m pre-trained_models/pm_dnn_model_5x5.h5 -s -v pre-trained_models/pm_dnn_value_model_5x5.h5