API para desenvolvimento de agentes de póquer online

• Autor(a) principal: Carlos Eduardo Mesquita Frias
• Data de Publicação: 2014
• Tipo de documento: Dissertação
• Idioma: por
• Título da fonte: Repositório Científico de Acesso Aberto de Portugal (Repositórios Cientìficos)
• Texto Completo: https://hdl.handle.net/10216/73494
• topic: Engenharia electrotécnica, electrónica e informática; Electrical engineering, Electronic engineering, Information engineering

The creation of agents that simulate human player's behaviour for games that are based on hidden information and non-deterministic nature is a very active area on Artificial Intelligence research domain. The Texas Hold'em variant of poker provides a context for studying the effectiveness of various implementation techniques and theories because of the properties it possesses such as the need to deal with hidden information and stochasticity. The most effective way to test, evaluate and compare methodologies used in the creation of poker agents would be the ability to put those agents playing against humans on an equal footing in a online playing environment at real time and with the same time and money constraints. The focus of this dissertation follows this idea, the creation of an API that provides the agents the necessary structures to play poker online, providing the state of the game as input to the agent and making the moves according to the output received from that agent. With the study of the state of the art, some approaches and techniques of computer vision, could form the basis and provide promising hopes on achieving the intended goals.