Artificial Intelligence with Graph Neural Networks Applied to a Risk-like Board Game

While AI has been successfully applied to many board games, such as chess and Go, most research is confined to a single board and is inflexible to topological changes. Contrarily, this research develops an AI agent, referred to as GG-net, to play an online strategy game based on the classic board game Risk, which is played on a wide variety of irregularly shaped maps. Prior research has struggled to create an effective AI for Risk-like games due to the immense branching factor. The most successful attempts tended to rely on manually restricting the AI's set of actions and providing the AI with handcrafted features. GG-net uses no human knowledge, instead, it relies on a genetic algorithm combined with a graph neural network. Together, these methods allow GG-net to overcome the high branching factor and generalize across a multitude of maps. GG-net appears to be a strong opponent on both small and medium maps, however, on large maps with hundreds of territories, inefficiencies become more significant and GG-net struggles against the rule-based agents.