Agent Model
How to Operate Agent Model : Basic
Click on the link, which will take you to AgentCubes. If you only want to
view the model as is, click ‘Play’. If you wish to play around with the
simulation settings and place your own agents, click on ‘Design.’ Click on the
Play button to begin the model. Click on the red square to stop the model.
Click on the triangle with the bar to observe the model passing one time step.
If you are on Design mode, you can click the pencil to place agents into the
model, and the eraser tool to erase agents, and the grid tool to place a large
amount of agents in a rectangle. Sliding the bar left towards the turtle will
slow down the model, while sliding the bar right toward the rabbit will speed
up the model. Click on the undo button to reset the model to its original state.
How to Operate Agent Model : Simulation Properties
To control some of the simulation properties that make up the model,
click on the gear bar on the left and then click on ‘Simulation Properties.’
A table should pop up where you are able to input the different values of the
four properties: hunger, death pheromone strength, follower pheromone
strength, and ant frequency. Either use the + and - buttons to increment the
values or manually input the values yourself to change how the model works
when the constants are changed.
What software do you need to open this model?
You can view the AgentCubes model without any software- it is free and online!
How model operates on bigger scales
This model works best on larger scales and bigger worlds. It makes the model
more realistic, for larger ant trails and time to scatter. In smaller scales,
the ants and predators don’t have enough space to roam and will encounter
each other very quickly. It increases the probability of complete annihilation
of a species and doesn’t provide useful information.
What did you do to try to break the model?
Like previously written, small scales breaks the model. In addition, placing
too many ant hills next to eachother messes with the diffusion of the
pheromones, because the ants themselves don’t have a value of the pheromone.
Analysis of Data found from Model
The model helped further my understanding of the effects of releasing ant
pheromones and the predator-prey dynamic between ants and frogs. With just
the default parameters of the model, you could observe that the pheromones
are vital for ants to communicate and find prey, and if a leader ant is
misled the whole trail could be marching in the wrong direction. With changing
the simulation properties of the model, specifically, the randomness in this
system also will change the outlook of the model every single run through,
whether through the direction of the pheromone trail, or the likelihood of an
ant mob overcoming a predator. Since all the agents move randomly in the world,
the chances of drastic changes to the population are possible, just like in
real life, but not likely. This is due to the fact that the ants are directed
by a pheromone trail and will follow it.