by Marcin Szubert, Paweł Liskowski, Wojciech Jaśkowski, Krzysztof Krawiec
Abstract:
In evolutionary learning of game-playing strategies, fitness evaluation is based on playing games with certain opponents. In this paper we investigate how the performance of these opponents and the way they are chosen influence the efficiency of learning. For this purpose we introduce a simple method for shaping the fitness function by sampling the opponents from a biased performance distribution. We compare the shaped function with existing fitness evaluation approaches that sample the opponents from an unbiased performance distribution or from a coevolving population. In an extensive computational experiment we employ these methods to learn Othello strategies and assess both the absolute and relative performance of the elaborated players. The results demonstrate the superiority of the shaping approach, and can be explained by means of performance profiles, an analytical tool that evaluate the evolved strategies using a range of variably skilled opponents.
Reference:
Shaping Fitness Function for Evolutionary Learning of Game Strategies (Marcin Szubert, Paweł Liskowski, Wojciech Jaśkowski, Krzysztof Krawiec), In GECCO’13: Proceedings of the 15th annual conference on Genetic and Evolutionary Computation (Christian Blum, ed.), ACM, 2013.
Bibtex Entry:
@InProceedings{Szubert2013shaping,
Title = {Shaping Fitness Function for Evolutionary Learning of Game Strategies},
Author = {Marcin Szubert and Pawe{ł} Liskowski and Wojciech Jaśkowski and Krzysztof Krawiec},
Booktitle = {GECCO'13: Proceedings of the 15th annual conference on Genetic and Evolutionary Computation},
Year = {2013},
Address = {Amsterdam, The Netherlands},
Editor = {Christian Blum},
Month = {July},
Pages = {1149--1156},
Publisher = {ACM},
Abstract = {In evolutionary learning of game-playing strategies, fitness evaluation is based on playing games with certain opponents. In this paper we investigate how the performance of these opponents and the way they are chosen influence the efficiency of learning. For this purpose we introduce a simple method for shaping the fitness function by sampling the opponents from a biased performance distribution. We compare the shaped function with existing fitness evaluation approaches that sample the opponents from an unbiased performance distribution or from a coevolving population. In an extensive computational experiment we employ these methods to learn Othello strategies and assess both the absolute and relative performance of the elaborated players. The results demonstrate the superiority of the shaping approach, and can be explained by means of performance profiles, an analytical tool that evaluate the evolved strategies using a range of variably skilled opponents.},
Url = {http://www.cs.put.poznan.pl/wjaskowski/pub/papers/szubert2013shaping.pdf}
}