This isn't exactly a demo. It's a sequence of examples that start with simple enemy AI and show how to add features of increasing complexity.
The package includes 15 examples. Each example shows how to implement a distinct feature useful for enemy AI. The concepts are explained in detail.
The examples cover things like:
• How to manage AI loops efficiently with thousands of mobs. (NEW!)
• How to create mobs that attack only when provoked and mobs that don't.
• How to use a simple state machine to create mobs that have a chance to run away when they are weak.
• How to create mobs that call for help so their allies will join them in combat.
• How to evaluate the utility of a situation and various actions to determine which action yields the largest increase in utility.
• How to couple utility functions with state machines to create mobs that can decide what action to perform and execute a multi-step action.
• How to make mobs run away to good hiding spots when they get weak.
• How to organize your AI code so that it doesn't get out of hand too quickly.
The package also includes ten exercises you can try to test your understanding of the concepts (complete solutions are provided for three of them).
Check out this blog post for instructions on how to run the examples.
Jun 27 2010, 9:00 pm