Publication: Entrenando bots para juegos mediante aprendizaje por refuerzo basado en casos
Loading...
Official URL
Full text at PDC
Publication Date
2017
Authors
Advisors (or tutors)
Editors
Journal Title
Journal ISSN
Volume Title
Publisher
Citations
Exportar
Abstract
Los videojuegos son un interesante campo de estudio para muchos investigadores de inteligencia artificial, dado que multitud de algoritmos distintos pueden ser estudiados y probados con ellos, y luego estos pueden ser aplicados a muchas otras situaciones. En este proyecto se utilizan principios de razonamiento basado en casos y aprendizaje por refuerzo para entrenar bots y que jueguen a Ms.Pac-Man. En concreto se utiliza el algoritmo Qlearning reemplazando la tabla Q y los estados por una base de casos. El uso de casos permite lidiar con una completa representación del estado del juego, y utilizar conocimiento experto sobre el dominio del juego tanto en la recuperación de casos como en la adaptación de soluciones.
Video games are an interesting field of study for many artificial intelligence researchers, since many different AI methods can be studied and tested with them, and later those investigations can be applied to many other situations. In this project case based reasoning and reinforcement learning principles are used to train bots to play the Ms. Pac-Man vs. Ghosts game. In particular, we use the well-known Q-learning algorithm but replacing the Q-table and states with a case base. The use of cases allows us to deal with rich game state representation and inject domain knowledge in both the retrieval and the adaptation stages.
Video games are an interesting field of study for many artificial intelligence researchers, since many different AI methods can be studied and tested with them, and later those investigations can be applied to many other situations. In this project case based reasoning and reinforcement learning principles are used to train bots to play the Ms. Pac-Man vs. Ghosts game. In particular, we use the well-known Q-learning algorithm but replacing the Q-table and states with a case base. The use of cases allows us to deal with rich game state representation and inject domain knowledge in both the retrieval and the adaptation stages.
Description
Trabajo de Fin de Grado en Ingeniería Informática y Matemáticas (Universidad Complutense, Facultad de Informática, curso 2016/2017)