The EmbeddedHunter Algorithm for Large-Scale Black-Box/Derivative-Free Optimization

This repository hosts the code for the EmbeddedHunter algorithm for large-scale black-box optimization proposed in the Embedded Bandits for Large-Scale Black-Box Optimization [paper] (https://arxiv.org/abs/1611.08773), which is accepted at the thirty-first AAAI conference on artificial intelligence ([AAAI-17] (http://www.aaai.org/Conferences/AAAI/2017/aaai17call.php)). Poster can be found [here] (https://github.com/ash-aldujaili/eh-lsopt/raw/master/aaai-poster.pdf).

Contents

1. Embedded Bandits for Large-Scale Black-Box Optimization.ipynb : a notebook for a quick demonstration.
2. run_aaai_demo.py: a python script for running AAAI’17 paper experiments.
3. Benchmark.py: a python module for the large-scale problems used.
4. Algorithms.py: a python module implementing RESOO, SRESOO, and EmbeddedHunter.
5. *.pickle: a collection of pickle files saving the outcome of run_aaai_demo.py
6. plot-table.tex: a tex file to reproduce Figure 1 of the AAAI’17 paper.

To run all the experiments (it will take around 6 days), cd to the directory and execute the following

>>python run_aaai_demo.py -1

Since all the *.pickle files are there, the above command will just create a folder, which has all the experiments’ results in a list of tabulated files. With these files at hand, you can now compile plot-table.tex to create the pdf version of AAAI’17 paper’s Figure 1.

Citation

If you write a scientific paper describing research that made use of this code, please consider citing the following paper:

@inproceedings{ash2017eh,
  title={Embedded Bandits for Large-Scale Black-Box Optimization},
  author={Abdullah Al-Dujaili and S. Suresh},
  booktitle={Thirty-First AAAI Conference on Artificial Intelligence (AAAI'17},
  year={2017}
}