Rolling horizon co-evolution in two-player general video game playing
- Artificial Intelligence for General Video Game Playing (GVGP) is challenging not only because agents must adapt to a range of different games, but they must also make decisions within the time constraints of real-time video games. The General Video Game Artificial Intelligence framework (GVGAI) is a popular framework for GVGP. It features a two-player track where two agents play a game together, either competitively or cooperatively, which poses the additional challenge of considering another player. Commonly, agents only consider their own moves in these two-player games. In this paper we discuss and assess Rolling Horizon Co-evolutionary Planning (a modification to Rolling Horizon Evolutionary Algorithms) for two player GVGAI. We present experimental results on its effectiveness against other agents playing GVGAI games and show that co-evolution can improve results compared to a RHEA agent.
Author: | Charles Ringer, Cristiana Pacheco, Georgiana Cristina DobreGND, Diego Perez-Liebana |
---|---|
Frontdoor URL | https://opus.bibliothek.uni-augsburg.de/opus4/112658 |
ISBN: | 978-1-7138-2942-3OPAC |
Parent Title (English): | AISB Convention 2021: Communication and Conversations, 7-9 April 2021, online |
Publisher: | Society for the Study of Artificial Intelligence and Simulation of Behaviour |
Place of publication: | Bath |
Type: | Conference Proceeding |
Language: | English |
Year of first Publication: | 2021 |
Release Date: | 2024/04/23 |
First Page: | 18 |
Last Page: | 23 |
Institutes: | Fakultät für Angewandte Informatik |
Fakultät für Angewandte Informatik / Institut für Informatik | |
Fakultät für Angewandte Informatik / Institut für Informatik / Lehrstuhl für Menschzentrierte Künstliche Intelligenz | |
Dewey Decimal Classification: | 0 Informatik, Informationswissenschaft, allgemeine Werke / 00 Informatik, Wissen, Systeme / 000 Informatik, Informationswissenschaft, allgemeine Werke |