In 1986, Craig Reynolds created a well-regarded AI program called Boids (a combination of bird and droids). This program created a fascinating bird-like flocking behavior, where little triangles moved around the screen in ways that remind the observer of flocking birds or fish. When the environment had obstacles, the boids broke up to steer around the obstacles and rejoin later. A collision between two flocks will usually end up in the flocks joining up and moving on. The Boids algorithm is an implementation of autonomous agents for AI. Each individual boid makes decisions based on a few simple rules and its immediate environment. That results in what is called emergent behavior, which is behavior that looks as if it was designed from the top down, but is not. The irony is that a top-down-implemented AI frequently looks less intelligent than allowing the individual agents to make their own decisions. It's a little like the old Soviet top-down command-and-control economy, versus a capitalist economy where individuals make decisions based on their immediate environment. In games, as in economics, you can also have a mixed system, where a top-down AI can send messages to autonomous agents giving them new goals or instructions. In the game we are writing, we have a single enemy spaceship, so the decision to manage AI from the top down or through autonomous agents does not really matter much, but because you may choose to expand the game in the future to support multiple enemies and their AIs, our agent will manage itself autonomously.