Asymmetric Co-Evolution for Imperfect-Information Zero-Sum Games

Abstract

We present an asymmetric co-evolutionary learning algorithm for imperfect-information zero-sum games. This algorithm is designed so that the fitness of the individual agents is calculated in a way that is compatible with the goal of game-theoretic optimality. This compatibility has been somewhat lacking in previous co-evolutionary approaches, as these have often depended on unwarranted assumptions about the absolute and relative strength of players. Our algorithm design is tested on a game for which the optimal strategy is known, and is seen to work well.

Cite

Text

Halck and Dahl. "Asymmetric Co-Evolution for Imperfect-Information Zero-Sum Games." European Conference on Machine Learning, 2000. doi:10.1007/3-540-45164-1_18

Markdown

[Halck and Dahl. "Asymmetric Co-Evolution for Imperfect-Information Zero-Sum Games." European Conference on Machine Learning, 2000.](https://mlanthology.org/ecmlpkdd/2000/halck2000ecml-asymmetric/) doi:10.1007/3-540-45164-1_18

BibTeX

@inproceedings{halck2000ecml-asymmetric,
  title     = {{Asymmetric Co-Evolution for Imperfect-Information Zero-Sum Games}},
  author    = {Halck, Ole Martin and Dahl, Fredrik A.},
  booktitle = {European Conference on Machine Learning},
  year      = {2000},
  pages     = {171-182},
  doi       = {10.1007/3-540-45164-1_18},
  url       = {https://mlanthology.org/ecmlpkdd/2000/halck2000ecml-asymmetric/}
}