Liang Xin, Wen Song, Zhiguang Cao, and Jie Zhang. Step-wise Deep Learning Models for Solving Routing Problems. IEEE Transactions on Industrial Informatics (TII), 2020. (In Press) [https://ieeexplore.ieee.org/document/9226142]
To download the pretrained models and data
python download.py;
tar -zxvf data.tar.gz;
cd SWTAM; tar -zxvf SWTAM_pretrained;
cd ../SWTAM; tar -zxvf ASWTAM_pretrained;
cd ..;To train the SW-TAM models for tsp and cvrp, respectively
python run.py --graph_size 20 --baseline rollout --run_name 'tsp20_rollout' --val_dataset ../data/tsp20.pkl
python run.py --graph_size 20 --baseline rollout --run_name 'vrp20_rollout' --val_dataset ../data/vrp20.pkl --problem="cvrp"To eval with pretrained models for tsp and cvrp, respectively
python3 run.py --graph_size 20 --load_path pretrained/tsp_20/epoch-99.pt --val_dataset ../data/tsp20.pkl --eval_only
python3 run.py --graph_size 20 --load_path pretrained/cvrp_20/epoch-99.pt --val_dataset ../data/vrp20.pkl --problem="cvrp" --eval_onlyTo train ASW-TAM;
python run.py --graph_size 20 --baseline rollout --run_name 'tsp20_rollout' --val_dataset ../data/tsp20.pklTo eval with pretrained models
python3 eval.py ../data/tsp20.pkl --model pretrained/tsp_100/epoch-99.pt --decode_strategy greedyTo train SW-PtrNet
python -u main.py --hidden_dim=256 --log_step=1000 --checkpoint_secs=30000To eval with pretrained_models
python -u main.py --hidden_dim=256 --log_step=1000 --checkpoint_secs=30000 --load_path pretrained/tsp_2019-08-19_14-04-45/model.ckpt-18000000 --is_train=FalseSWTAM and ASWTAM are developed based on https://github.com/wouterkool/attention-learn-to-route SWPtrNet are developed based on https://github.com/devsisters/pointer-network-tensorflow
React
Typescript
Django
Laravel
Facebook
Microsoft
Google
Alibaba
D3
Tencent