Sometimes for your games, you might want to create crowd AIs that act in a very unique way, and instead of using an existing AI plugin, you might want to create a crowd AI from scratch. This would be done in the case of an ANT-Op AI. As we mentioned before, ant behavior is a popular topic in AI research and computer science in general. Initially, ants work independently and give off pheromones that are sensed by other ants to communicate messages. For example, when ants start searching for food, they give off pheromones as they search. When they find food, they give off different pheromones as they bring it back to the colony, which directs the next ants when searching for food. ANT-Op, by Gray Lake Studios, is available on Unity Asset Store for $75, and it simulates this ant food search process. Unlike the other AI plugins in this book, Ant-Op isn't really designed to be brought into an existing game; it's more of a technical demonstration that is a simulation you can use to see interesting AI at work and hopefully use it to inspire complex AI designs for your games. To start the demo, import ANT-Op and double-click on Test Scene to open it. The scene will initially be blank, but if you start the demo, you can see the simulation start. In the following screenshot, you can see ANT-Op in action:

You can see ANT-Op simulating an ant colony. The lines represent pheromones from the ants.

The options on the right are different settings for the simulation. You can play by changing the values, and these can be saved to an XML file for later reloading. Running the simulation provides a complex visualization, which you can see in the following screenshot:

Again, this probably isn't something you would bring into your game as is; it's to generate ideas for your AI. The source files can be reviewed for ideas on creating your own AI. Don't underestimate what you can learn by looking at AI simulations like this.