A Neurogenetic Framework for Problem Solution Applied to a Game Situation

This paper presents Neurogenetic Connect Four (NGC4), an evolutionary connectionist game solution, which, in a novel attempt, combines evolution and artificial neural networks in order to automatically learn a game of moderate difficulty. Besides the complete system design, underlying principles are introduced and expectations are discussed. It is firmly estimated that the algorithm might not only give way to a truly automatic solution of game situations, but also have the potential to offer a starting point for automated problem solution systems outside toy domain