Multi-objective AutoML for Enhanced Vision-based Structural Health Monitoring: A New Reinforcement-Learning-based Metaheuristic Approach

This repository contains the used data and algorithm proposed in our research paper [1].

The repository is organized as follows:

• "Experiment_1": Contains the data and code used for the first experiment.
• "Experiment_2": Includes the data and code used for the second experiment.

Inside each folder, you will find a Jupyter notebook containing the algorithm code, as well as a "content" folder that contains the corresponding data split into training, validation, and testing subfolders.

To run the codes, it is recommended to use a machine with a GPU and install the TensorFlow 2.10.1 library for Python. The code has been successfully tested with Python 3.9.18.

We appreciate your cooperation in acknowledging the original work [1] and giving appropriate credit to the authors. For any feedback or inquiries regarding this code, please contact Armin Dadras Eslamlou at armin.dadras.eslamlou@gmail.com.

Reference:

[1] Armin Dadras Eslamlou, Shiping Huang, Reinforcement learning for multi-objective AutoML in vision-based structural health monitoring, Automation in Construction, Volume 166, 2024, https://doi.org/10.1016/j.autcon.2024.105593. (https://www.sciencedirect.com/science/article/pii/S0926580524003297)