Requires python 3.11 and PlatEMO v4.2 (https://github.com/BIMK/PlatEMO).
Clone this repository and install the packages specified in requirements.txt
git clone https://github.com/Anomaly33/RLXBench.git
cd RLXBench
pip install -r requirements.txt
For the reacher environment, you'll need to install pybullet-gym from https://github.com/benelot/pybullet-gym
git clone https://github.com/benelot/pybullet-gym.git
cd pybullet-gym
pip install -e .
you should then copy the DRL folder to the PlatEMO multi-objective problem directory, the mat_eval_env.py to the main PlatEMO directory (the one with the platemo.m file), and the HV_rl.m file to the Metric directory
Path related information
pyenv("Version",'C:\Users\ecis\anaconda3\envs\RL_Bench\python.exe')
create metric files for HV and Sparsity
the following environments are currently implemented
| Environment ID | Description |
|---|---|
| DRL1 | Deep Sea Treasure |
| DRL2 | Deep Sea Treasure Concave |
| DRL3 | Fruit Tree |
| DRL4 | Four Room |
| DRL5 | Fishwood |
| DRL6 | Minecart |
| DRL7 | Ant |
| DRL8 | Hopper |
| DRL9 | Half Cheetah |
| DRL10 | Reacher |
| DRL11 | Walker 2D |
| DRL12 | Lunar Lander |
If you find this code or project helpful in your research, please cite our paper:
@article{ajani2024deep,
title={Deep reinforcement learning as multiobjective optimization benchmarks: Problem formulation and performance assessment},
author={Ajani, Oladayo S and Ivan, Dzeuban Fenyom and Darlan, Daison and Suganthan, PN and Gao, Kaizhou and Mallipeddi, Rammohan},
journal={Swarm and Evolutionary Computation},
volume={90},
pages={101692},
year={2024},
publisher={Elsevier}
}