The environment will play an important role when designing the AI. First and foremost, there are some additional restrictions that relate to the problem:
The structure of the environment has a direct effect on the movement.
Movement simulation is always handled discretely in time, so the AI gets limited opportunities to move.
There is no way of gathering perfect information about the world.
We deliberated on some rough guidelines for the application stages:
We can consider the world as "mostly" 2D thanks, to gravity.
It's important to consider how to handle other characters in the environment.
It's possible to use (seemingly) continuous movement in space, but discretization is a possible alternative.
Finally, we identified some assumptions about the platform to use in the specification phase:
The low-level details of animation (for instance, locomotion) are abstracted out from the AI.
The physics engine is in charge of conflict resolution between the players' movement and the environment.
It's possible to extract information about the environment for the simulation.
After the practical interlude, the next chapter considers the art of navigation, and how it can be reproduced artificially.
A small game level is sufficient to illustrate the ideas in this chapter. An example animat called Dr. Spin can be found on the web site at http://AiGameDev.com/, along with a guide to run the binary demo. Dr. Spin's only behaviors are to move and turn, often bumping into walls and getting stuck in corners. These situations demonstrate the importance of the world simulation and the structure on the movement.