ABSTRACT
While there are many different architectures for an AI programmer to pick from, behavior trees are one of the most popular algorithms for implementing NPC action selection in games due to their simplicity to code and use. They are quick to implement from scratch and can be extended to add additional features or provide game-specific functionality as needed. While not as simple as a finite-state machine, they are still simple enough to be easily debugged and designed by other team members as well, making them appropriate to use on games with a large team implementing the behaviors.
