Evolutionary neural networks for Non-Player Characters in Quake III

Modelagem De Comportamento De Jogadores Por Meio De Agrupamento Online E Detecção De Novidades

“…Em geral, utilizam técnicas de aprendizado supervisionado em métricas de jogos para a obtenção de modelos de decisão, sendo necessário para isso a criação de conjuntos de dados contendo informações rotuladas daquilo que se deseja aprender [3,4]. Existem ainda trabalhos que utilizam abordagens evolutivas como mecanismo de aprendizado aplicado aos algoritmos de AM [5,6,7]. Abordagens não supervisionadas também têm sido exploradas para a definição de grupos que representam estilos de jogo [8,9].…”

Fuzzy Tactics: A scripting game that leverages fuzzy logic as an engaging game mechanic

“…Fuzzy-logic controlled bots have been released for the game ( van Waveren, 2001), with weapon and item selection controlled by fuzzy decision making. The fuzzy bots have shown interesting performance and they are now taken as the basis for comparison when more advanced techniques are used (Prieditis & Dalal, 2006;Westra & Dignum, 2009). Pinto and Alvares (2006) model fuzzy sensors as input to extended behavior networks in Unreal Tournament (Epic, 1999), another famous first person shooter, while Acampora (2010) uses timed automata and fuzzy controllers to model emotions for bots in a Unreal Tournament 2004 (Epic, 2003) match.…”

Creating Believable and Effective AI Agents for Games and Simulations

“…Referenced in (Choi et al, 2007) as a testbed for a believable agent, built with the help of general-purpose cognitive architecture called ICARUS. Also used in experiments with neural network-based AI by Westra and Dignum (2009), and with genetic algorithm-based AI strategy optimization by Liaw, Wang, Tsai, Ko, & Hao (2013). Since the engine's source code is freely available now under GNU license, some researchers implement new algorithms directly in the code of existing game bots (El Rhalibi & Merabti, 2008).…”