Top AI Repos — open-source AI, indexed and scored
Top AI Repos tracks AI repositories on GitHub and answers two different questions about each one: is it moving right now, and would you bet a product on it.
Top AI Repos tracks AI repositories on GitHub and answers two different questions about each one: is it moving right now, and would you bet a product on it.
Graph Neural Networks for Decentralized Path Planning
| Date | Stars |
|---|---|
| 2026-07-31 | 257 |
| 2026-08-03 | 257 |
| 2026-08-04 | 257 |
| 2026-08-13 | 257 |
| 2026-08-14 | 257 |
| 2026-08-18 | 257 |
| 2026-08-28 | 256 |
| 2026-08-31 | 256 |
| 2026-09-01 | 256 |
| 2026-09-06 | 256 |
| 2026-09-11 | 255 |
| 2026-09-20 | 255 |
Today
— stars today
This week
— stars this week
This month
-2 stars this month
Momentum
0.0
growth rate 0.00%/day
# News 2021
We created a new [repo: magat_pathplanning](https://github.com/proroklab/magat_pathplanning) that integrated this repo and MAGAT (RAL2021) with several major updates that provide training speed-up, improvements to simulator, rework of code structure, and clearer comments.
We highly recommend to use the new repository for replicating and experimenting the GNN path-planner in this page.
# PyTorch Project for Graph Neural Network based MAPF
Code accompanying the paper
[Graph Neural Networks for Decentralized Multi-Robot Path Planning](https://arxiv.org/abs/1912.06095)
from Qingbiao Li (1), Fernando Gama (2), Alejandro Ribeiro (2), Amanda Prorok (1) at University of Cambridge (1) and at University of Pennsylvania (2).
### Table of Contents:
<!-- Table of contents generated generated by http://tableofcontent.eu -->
- [Graph MAPF project](#pytorch-project-template)
- [Project Diagram](#template-class-diagram)
- [Framework Structure](#repo-structure)
- [How to use this repo](#use-repos)
- [Requirements](#requirements)
- [License](#license)
- [Citation](#Citation)
### Project Diagram:
### Framework Structure:
The repo has the following structure:
```
├── agents (overall framework for training and testing)
| └── base.py
| └── decentralplannerlocal.py
| └── (DCP)
| └── decentralplannerlocal_OnlineExpert.py
| └── (DCP with onlin expert mechanism)
|
├── configs (set up key parameters for training and inference stage,)
| └── dcp_ECBS.json
| └── dcp_onlineExpert.json
|
├── dataloader (load data for training)
| └── Dataloader_dcplocal_notTF_onlineExpert.py
|
├── graphs
| └── models (model including CNN -> GNN -> MLP)
| | |
| | └── decentralplanner.py
| |
| └── losses
| └── cross_entropy.py
|
├── utils
| |
| └── assets
| | └── dataTools.py
| | └── graphML.py
| | └── graphTools.py
| |
| └── multirobotsim_dcenlocal.py
| └── (simulator for dencentral agents)
| └── multirobotsim_dcenlocal_onlineExpert.py
| └── (simulator for dencentral agents with online expert mechanism, where failure is saved.)
| └── visualize.py
| └── (visualize the predicted path with communcation link.)
| └── visualize_expertAlg.py
| └── (visualize the ground truth path.)
| └── metrics.py
| └── (Record stastics during inference stage.)
| └── config.py
|
├── offlineExpert
| |
| └── CasesSolver.py
| └── 1, (# generate map) Randomly generate map with customized obstacle density and obstacle,
| └── 2. (# case under a map)
| └── At each specific map, generate random pairs of start and goal position for each agents.
| └── 3. (for given case) Apply expert algorithm to compute solution.
| |
| └── DataGen_Transformer.py
| └── (Transform the solution into specific data format that ready to be loaded by dataloader.
| └── including: map, input tensor wiith each agents paths, GSO.)
|
├── onlineExpert
| |
| └── ECBS_onlineExpert.py
| └── (Apply expert algorithm to compute solution for failture cases recorded during training process.)
| |
| └── DataTransformer_local_onlineExpert.py
| └── (Transform the solution into specific data format, and then merged into offline dataset.)
|
├── experiments
|
├── data
|
├── statistic_analysis
| └── (Fig.3.) result_analysis_errorbar.py
| └── (Fig.4.) result_analysis_generalization_colormap.py
| └── (Fig.5.) result_analysis_hist_impact_3K.py
|
└── main.py
```
### Requirements:
```
easydict>=1.7
matplotlib>=3.1.2
numpy>=1.14.5
Pillow>=5.2.0
scikit-image>=0.14.0
scikit-learn>=0.19.1
scipy>=1.1.0
tensorboardX>=1.2
torch>=1.1.0
torchvision>=0.3.0
```
### How to use this repo:
#### Test trained network, for exmaple DCP OE - K=3
1. [Download](https://drive.google.com/Excerpt of 5,800 characters
Read on GitHubWould you bet a product on this? Bounded 0–100 and slow moving.
matched fp:cd2db1563cfbdc88, llm:Repository description: 'Graph Neural Networks for Decentralized Path Planning' (Python).
matched fp:cd2db1563cfbdc88, llm:Repository description: 'Graph Neural Networks for Decentralized Path Planning' (Python).
matched fp:cd2db1563cfbdc88, llm:Repository description: 'Graph Neural Networks for Decentralized Path Planning' (Python).