Thinklab-SJTU / GAL-VNE

Official Implementation of KDD 2023 paper: "GAL-VNE: Solving the VNE Problem with Global Reinforcement Learning and Local One-Shot Neural Prediction"

GAL-VNE

Official Implementation of KDD 2023 paper:
"GAL-VNE: Solving the VNE Problem with Global Reinforcement Learning and Local One-Shot Neural Prediction"

VNE Problem

The NP-hard combinatorial Virtual Network Embedding (VNE) Problem refers to finding the node and edge mapping between a virtual network (VN) and the substrate network (SN).

GAL-VNE

We propose Global-And-Local VNE framework (GAL-VNE) by a pretrain-then-finetune training paradigm.
For more details, please see this project.

Requirements

easydict==1.10    
gym==0.26.2    
Jinja2==3.1.2    
matplotlib==3.8.3    
networkx==3.0    
numpy==1.23.5    
ortools==9.4.1874    
pandas==2.2.2    
PyYAML==6.0    
scikit_learn==1.2.2    
scipy==1.13.0    
torch==1.12.1    
torch_geometric==2.2.0    
tqdm==4.64.1    

Specify configurations

Customize "--pn_setting_path" and "--vns_setting_path".
The default setting is:

--pn_setting_path=settings/pn_setting/pn_setting.yaml --vns_setting_path=settings/vns_setting/vns_setting.yaml

Parameters	SN	VN
Number of SN	1	2000
Number of Nodes	100	$2 \sim 10$
Number of Links	500	$Random(0.5)$
Node CPU (uniform dist.)	$50 \sim 100$	$0 \sim 50$
Bandwidth (uniform dist.)	$50 \sim 100$	$0 \sim 50$
Arrival Time (Possion dist.)	-	$Poisson(0.1)$
Life Time (exponential dist.)	-	$Exp(500) $

The setting VN dataset during training is specified in "--vns_sub_setting_path".

Generate data

Run any solver with:

--generate_vn --generate_pn

Remind that generate the dataset once, and keep fixed afterwards. The default dataset is put in "dataset/" folder.

Reproduce the results

bash run.sh

Citation

If you find our work useful in your research, please consider citing:

@inproceedings{geng2023galvne,
    author = {Geng, Haoyu and Wang, Runzhong and Wu, Fei and Yan, Junchi},
    title = {GAL-VNE: Solving the VNE Problem with Global Reinforcement Learning and Local One-Shot Neural Prediction},
    year = {2023},
    isbn = {9798400701030},
    publisher = {Association for Computing Machinery},
    address = {New York, NY, USA},
    url = {https://doi.org/10.1145/3580305.3599358},
    doi = {10.1145/3580305.3599358},
    booktitle = {Proceedings of the 29th ACM SIGKDD Conference on Knowledge Discovery and Data Mining},
    pages = {531–543},
    numpages = {13},
    keywords = {combinatorial optimization, virtual network embedding, reinforcement learning},
    location = {Long Beach, CA, USA},
    series = {KDD '23}
}

Acknowledgement

We sincerely thank the repository virne for their well-implemented pipeline upon which we build our codebase.