{"id":25387170,"url":"https://github.com/ntt-dkiku/evrp-eps","last_synced_at":"2026-03-09T10:31:36.251Z","repository":{"id":216260774,"uuid":"740860160","full_name":"ntt-dkiku/evrp-eps","owner":"ntt-dkiku","description":"The official implementation of \"Electric Vehicle Routing for Emergency Power Supply with Deep Reinforcement Learning\" (AAMAS 2024, extended abstract).","archived":false,"fork":false,"pushed_at":"2024-04-08T02:49:07.000Z","size":25210,"stargazers_count":7,"open_issues_count":0,"forks_count":2,"subscribers_count":1,"default_branch":"main","last_synced_at":"2025-06-14T13:43:42.346Z","etag":null,"topics":["combinatorial-optimization","deep-reinforcement-learning","electric-vehicle-routing-problem","electric-vehicles","multi-agent-reinforcement-learning","multi-agent-systems","transformer","vehicle-routing-problem"],"latest_commit_sha":null,"homepage":"https://ntt-dkiku.github.io/rl-evrpeps/","language":"Jupyter Notebook","has_issues":true,"has_wiki":null,"has_pages":null,"mirror_url":null,"source_name":null,"license":"other","status":null,"scm":"git","pull_requests_enabled":true,"icon_url":"https://github.com/ntt-dkiku.png","metadata":{"files":{"readme":"README.md","changelog":null,"contributing":null,"funding":null,"license":"LICENSE","code_of_conduct":null,"threat_model":null,"audit":null,"citation":null,"codeowners":null,"security":null,"support":null,"governance":null,"roadmap":null,"authors":null,"dei":null,"publiccode":null,"codemeta":null}},"created_at":"2024-01-09T08:07:07.000Z","updated_at":"2025-04-10T15:17:57.000Z","dependencies_parsed_at":"2024-01-09T09:33:02.740Z","dependency_job_id":"e2150e41-105c-402d-87c3-9fa64d6a30c2","html_url":"https://github.com/ntt-dkiku/evrp-eps","commit_stats":null,"previous_names":["ntt-dkiku/evrp-eps"],"tags_count":0,"template":false,"template_full_name":null,"purl":"pkg:github/ntt-dkiku/evrp-eps","repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/ntt-dkiku%2Fevrp-eps","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/ntt-dkiku%2Fevrp-eps/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/ntt-dkiku%2Fevrp-eps/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/ntt-dkiku%2Fevrp-eps/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/ntt-dkiku","download_url":"https://codeload.github.com/ntt-dkiku/evrp-eps/tar.gz/refs/heads/main","sbom_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/ntt-dkiku%2Fevrp-eps/sbom","scorecard":null,"host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":286080680,"owners_count":30291807,"icon_url":"https://github.com/github.png","version":null,"created_at":"2022-05-30T11:31:42.601Z","updated_at":"2026-03-09T02:57:19.223Z","status":"ssl_error","status_checked_at":"2026-03-09T02:56:26.373Z","response_time":61,"last_error":"SSL_read: unexpected eof while reading","robots_txt_status":"success","robots_txt_updated_at":"2025-07-24T06:49:26.215Z","robots_txt_url":"https://github.com/robots.txt","online":false,"can_crawl_api":true,"host_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub","repositories_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories","repository_names_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repository_names","owners_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners"}},"keywords":["combinatorial-optimization","deep-reinforcement-learning","electric-vehicle-routing-problem","electric-vehicles","multi-agent-reinforcement-learning","multi-agent-systems","transformer","vehicle-routing-problem"],"created_at":"2025-02-15T11:34:09.691Z","updated_at":"2026-03-09T10:31:36.227Z","avatar_url":"https://github.com/ntt-dkiku.png","language":"Jupyter Notebook","funding_links":[],"categories":[],"sub_categories":[],"readme":"# Electric Vehicle Routing for Emergency Power Supply: Towards Telecom Base Station Relief\n\n\u003cp align=\"center\"\u003e\n  \u003ca href=\"https://ntt-dkiku.github.io/rl-evrpeps/\" target=\"_blank\"\u003e\u003cimg src=\"https://img.shields.io/badge/Project-page-blue\"\u003e\u003c/a\u003e\n  \u003ca href=\"https://arxiv.org/abs/2404.02448\" target=\"_blank\"\u003e\u003cimg src=\"https://img.shields.io/badge/arXiv-abs-red\"\u003e\u003c/a\u003e\n  \u003ca href=\"https://www.aamas2024-conference.auckland.ac.nz/\" target=\"_blank\"\u003e\u003cimg src=\"https://img.shields.io/badge/AAMAS-2024-green\"\u003e\u003c/a\u003e\n\u003c/p\u003e\n\nThis repo is the official implementation of Electric Vehicle Routing for Emergency Power Supply with Deep Reinforcement Learning (AAMAS 2024, extended abstract) and [Electric Vehicle Routing for Emergency Power Supply: Towards Telecom Base Station Relief](https://arxiv.org/abs/2404.02448) (arXiv preprint).\n\n\u003cdiv\u003e\u003cvideo autoplay loop controls src=\"https://github.com/ntt-dkiku/evrp-eps/assets/154794155/818b81f7-8e19-40ac-9934-70bc30f13e53\"\u003e\u003c/video\u003e\u003c/div\u003e\n\n## 📦 Python Environment\nWe recommend using Docker to construct the python environment. You can use the [Dockerfile](./Dockerfile) in this repository. \n```\ndocker build -t evrp-eps/evrp-eps:1.0 .\n``` \nYou can run code interactively with the following command (\u003c\u003e indicates a placeholder, which you should replace according to your settings.\").\n```\ndocker run -it --rm -v \u003c/path/to/clone/repo\u003e:/workspace/app --name evrp-eps -p \u003chost_port\u003e:\u003ccontainer_port\u003e --gpus all evrp-eps/evrp-eps:1.0 bash\n```\n\n## 🔧 Usage\n### 1. Generating synthetic data\nFirst of all, we generate synthetic datasets for training/validation/evaluation. We recommend more than 12.8M training samples.\nIf you want to change some parameters, check the other options with ```python generate_datasets.py -h```.\n```\npython generate_dataset.py --save_dir data/synthetic_data --type all --num_samples 1280000 10000 10000\n```\n\n### 2. Training\nWe train the RL model on the synthetic datasets. Check the other options with ```python train.py -h```.\n```\npython train.py --dataset_path data/synthetic_data/train_dataset.pkl --checkpoint_dir checkpoints/demo_model --batch_size 256 --vehicle_speed 41 --wait_time 0.5 --time_horizon 12 --gpu 0\n```\n\n### 3. Validation\nWe determine the best epoch evaluating the RL-model with greedy decoding on the validation split. Check the other options with ```python valid.py -h```.\n```\npython valid.py --model_dir checkpoints/demo_model --dataset_path data/synthetic_data/valid_dataset.pkl --gpu 0\n```\n\n### 4. Evaluation\nThe option ```--model_dir \u003ccheck_point_dir\u003e``` automatically selects the weights at the best epoch. You can also select a specific epoch by ```--model_path model_epoch\u003cepoch\u003e.pth```. Specify only one of the two. If you want to output the visualization, add the option ```--visualize_routes```. Check the other options with ```python eval.py -h```.\n```\npython eval.py --model_dir checkpoints/demo_model --dataset_path data/synthetic_data/eval_dataset.pkl --vehicle_speed 41 --wait_time 0.5 --time_horizon 12 --gpu 0\n```\n\n## 🧪 Reproducibility\nRegarding the synthetic datasets, you can reproduce our experimental results in [reproduce_results.ipynb](./reproduce_results.ipynb).  \nPlease take a glance at the content via nbviwer: [nbviwer-evrp-eps](https://nbviewer.org/github/ntt-dkiku/evrp-eps/blob/main/reproduce_results.ipynb).\nYou can open the Jupyter Notebook server with the following command inside the container, then access it from your browser on localhost.\n```\njupyter lab --allow-root --no-browser --ip=0.0.0.0 --port \u003ccontainer_port\u003e\n```\n\n## 🐞 Bug reports and questions\nIf you encounter a bug or have any questions, please post issues in this repo.\n\n## 📄 Licence\nOur code is licenced by NTT. Basically, the use of our code is limitted to research purposes. See [LICENSE](./LICENSE) for more details.\n\n## 🤝 Citation\nIf you find this work useful, please cite our paper as follows:\n```\n@misc{kikuta2024electric,\n      title={Electric Vehicle Routing Problem for Emergency Power Supply: Towards Telecom Base Station Relief}, \n      author={Daisuke Kikuta and Hiroki Ikeuchi and Kengo Tajiri and Yuta Toyama and Masaki Nakamura and Yuusuke Nakano},\n      year={2024},\n      eprint={2404.02448},\n      archivePrefix={arXiv},\n      primaryClass={math.OC}\n}\n```\n","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fntt-dkiku%2Fevrp-eps","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Fntt-dkiku%2Fevrp-eps","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fntt-dkiku%2Fevrp-eps/lists"}