rg223/eps

Evolutionary Preference Sampling for Pareto Set Learning

The overview of the EPS.

This project is a pytorch implementation of Evolutionary Preference Sampling for Pareto Set Learning

Installation

Requirements

Our provide the packages file of our environment (requirement.txt), you can using the following command to download the environment:

• pip install -r requirements.txt

Parameters

• problem_name: Testing problems. (0: Normal, 1:EPS strategy)
• scalar: Scalarization function (tch2: Tchebycheff function, tch1: Modified Tchebycheff function, cosmos: A Feature fusion method, hv1: hypervolume optimization)

Training

To train a model with GPU run:

• Using EPS strategy:

cd /projects/EPS
python train_net.py --problem_name 'dtlz7' --using_adapt 1

• Do not use EPS strategy:

cd /projects/EPS
python train_net.py --problem_name 'dtlz7' --using_adapt 0

Citation

Our paper can be accessed at https://arxiv.org/abs/2404.08414.

@article{ye2024evolutionary, title={Evolutionary Preference Sampling for Pareto Set Learning}, author={Ye, Rongguang and Chen, Longcan and Zhang, Jinyuan and Ishibuchi, Hisao}, journal={arXiv preprint arXiv:2404.08414}, year={2024} }