In this paper we show how combining fuzzy sets and reinforcement learning a winning agent can be created for the popular Pac-man game. Key elements are the classification of the state into a few fuzzy classes that makes the problem manageable. Pac-man policy is defined in terms of fuzzy actions that are defuzzified to produce the actual Pac-man move. A few heuristics allow making the Pac-man strategy very similar to the Human one. Ghosts agents, on their side, are endowed also with fuzzy behavior inspired by original design strategy. Performance of this Pac-man is shown to be superior to those of other AI-based Pac-man described in the literature.
Clever Pac-man / N.A. Borghese, C. Quadri, N. Rossini - In: Neural nets WIRN11 : proceedings of the 21st Italian workshop on neural nets / [a cura di] B. Apolloni, S. Bassis, A. Esposito, C.F. Morabito. - Amsterdam : IOS press, 2011. - ISBN 9781607509714. - pp. 11-19 (( Intervento presentato al 21. convegno Italian Workshop on Neural Networks (WIRN) tenutosi a Vietri sul mare, Italy nel 2011 [10.3233/978-1-60750-972-1-11].
Clever Pac-man
N.A. BorghesePrimo
;C. QuadriSecondo
;
2011
Abstract
In this paper we show how combining fuzzy sets and reinforcement learning a winning agent can be created for the popular Pac-man game. Key elements are the classification of the state into a few fuzzy classes that makes the problem manageable. Pac-man policy is defined in terms of fuzzy actions that are defuzzified to produce the actual Pac-man move. A few heuristics allow making the Pac-man strategy very similar to the Human one. Ghosts agents, on their side, are endowed also with fuzzy behavior inspired by original design strategy. Performance of this Pac-man is shown to be superior to those of other AI-based Pac-man described in the literature.
Pubblicazioni consigliate
I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
