The agent generation system allows for a very modular workflow, however there are some considerations that you should keep in mind.


The standard workflow for agent generation is:

  1. Objects (Geometry)
  2. Templates (Agent Info)
  3. Positioning (Layout)
  4. Generate (Output)

The nodes in the system are arranged so as to make this order apparent. In the add menu for the node tree, the groups of nodes are arranged in the order of which they should be used.


Objects (Geometry)

This is the actual geometry that will be duplicated and modified to create the pack of agents. These objects can be single or they can be grouped together.

Included in this group are modifiers that (among other things) allow you to randomize which objects are selected for generation.

Templates (Agent Info)

The template tools are used to connect the agents generated to the animation/simulation system. Most important is the Template node. This node does most of the dirty work and at least one is required in every tree.

Additionaly, the Template node takes the objects inputted and joined them into a virtual group that can be scaled, rotated, etc.

Positioning (Layout)

The positioning nodes are quite self-explanatory. These nodes are responsibe for duplicating the template(s) inputted and arranging them in the scene.

Half of the nodes place the templates and the other half (such as Obstacle and Ground) modify those placements. Make sure that the modifier nodes always come before the placement nodes.

Generate (Output)

The last node in the agent generation tree is always the Generate node. There should only be one Generate node in the tree.

Common Mistakes

  • One or more of the positioning modification nodes are placed after one or more of the positioning placement nodes

This causes the modification nodes to not have any effect at all.

Example Node Trees

Simplest Possible Setup

This node tree has the bare minimum number of nodes needed for agent generation.


Random Geometry Input

This node tree takes multiple pieces of geometry and randomly decides which one to use for each agent in the template.


Random Template Input

This node tree takes multiple templates and randomly decides which one to use for each position in the generation. This is different from the example above because it randomly choosed which virtual group of agents with its own node tree to use, not which geometry.


Formation Of Formations

This node tree demonstrates the power of the node-based system. It allows you to generate duplicates and groups of already generated agents. This example shows how a group of agents arranged in a formation can be arranged in another formation.