https://github.com/zagoli/cmapdcostregression
Predict the cost of a CMAPD assignment with XGBoost regression
https://github.com/zagoli/cmapdcostregression
multiagent-planning pathfinding regression xgboost
Last synced: 4 days ago
JSON representation
Predict the cost of a CMAPD assignment with XGBoost regression
- Host: GitHub
- URL: https://github.com/zagoli/cmapdcostregression
- Owner: zagoli
- License: mit
- Created: 2023-01-30T09:14:28.000Z (over 3 years ago)
- Default Branch: main
- Last Pushed: 2023-03-30T10:15:01.000Z (about 3 years ago)
- Last Synced: 2025-03-04T15:48:15.529Z (about 1 year ago)
- Topics: multiagent-planning, pathfinding, regression, xgboost
- Language: C++
- Homepage:
- Size: 6.16 MB
- Stars: 0
- Watchers: 1
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
# CMAPD solution cost regression
This code tries to predict the cost of a conflict free solution
for a CMAPD instance (with already assigned tasks) using regression.
The best $R^2$ score achieved so far is 0.97!
### Installation
Installation instructions for Linux and similar systems:
1. create a conda environment with the provided requirements file: `conda_env_requirements.txt`;
2. clone this repo;
3. activate the environment and navigate to `cmapdCostRegression/oracle`;
4. run `make`.
All done!
### Usage
How to train the model:
1. change the value of the configuration variables in `config.py` if needed;
2. run `python datasetGenerator.py` to generate the dataset based on the config variables;
3. run `python train.py` to train the model.
The trained model will also be saved for later usage.