Currently, the design of the model is such that an Agent is associated with only one parent organization. This organization is associated with the an agent when it is created. The Agent uses this information to determine which kinds of moves are appropriate and which are not.
It could be argued that an Agent should not this information provided, as it constrains the way an Agent can be used.
The design of this model acknowledges this problem in two ways:
Information extracted from the Organization (which is limited to information about the dimensions of the space), is private to the Agent class. Specific agent types have to depend on methods that provide relativistic information in order to make decisions.
By introducing an abstract Organization class that is intended to be a very general (as opposed to a standalone SDG class). The public methods and variables of the organization reveal minimal information. (Information that could inadvertently entangle agents and their landscape.)
Size is used for different things in different agents. It's a generic measure of (self) importance.
Agents can do a random walk in some direction. This parameter indicates how big the radius of a step can be.
Although agents have an absolute position, they only change it via offsets. Often it isn't realistic to have an agent know where it is. In this case, the space that the agents explore is a conceptual space, so it is silly to take the notion of absolute coordinates seriously -- if they knew what the range and density of the coordinates, there would be no point in exploring it. However, having absolute coordinates can be handy for display purposes. Be careful!