fast_map-elites

The fastest MAP-Elites implementation?

On a MacBook 2020, about0.5 s for 1M evaluations of the kinematic arm.

Why doing this?

• benchmarking new ideas in a few minutes of computation
• hyperparameter tuning
• online optimization

And it is good to have implementations that are not wasteful when possible!

Algorithms

• MAP-Elites: Mouret JB, Clune J. Illuminating search spaces by mapping elites. arXiv preprint arXiv:1504.04909. 2015 Apr 20.
• Using centroids: Vassiliades V, Chatzilygeroudis K, Mouret JB. Using centroidal voronoi tessellations to scale up the multidimensional archive of phenotypic elites algorithm. IEEE Transactions on Evolutionary Computation. 2017 Aug 3;22(4):623-30.
• Variation operator: Vassiliades V, Mouret JB. Discovering the elite hypervolume by leveraging interspecies correlation. InProceedings of the Genetic and Evolutionary Computation Conference 2018 Jul 2 (pp. 149-156).
• Bounce back operator: Nordmoen J, Nygaard TF, Samuelsen E, Glette K. On restricting real-valued genotypes in evolutionary algorithms. InInternational Conference on the Applications of Evolutionary Computation (Part of EvoStar) 2021 Apr 7 (pp. 3-16). Springer, Cham.

Compilation:

• clang++ -DUSE_BOOST -DUSE_TBB -ffast-math -O3 -march=native -std=c++14 -I /usr/local/include/eigen3 map_elites.cpp -o map_elites_2 -ltbb