Homework

MCB 419 Homework 9 (Spring 2008)

Questions appearing in red are to be answered in the hw09.txt file.

When you've finished the assignment, email your responses and a copy of your project file to mcb419@gmail.com with 'hw09' in the subject line.

The homework is due by 11:59PM on Tue, Apr 8.

Ant Pheromone Trails

AntTrails Improving foraging efficiency via a pheromone-based communication system.

In this project, you'll develop a communication system based on pheromones to improve ant foraging efficiency. The goal is for the ant colony to collect as much food as possible in the time alloted (50 s). Color code: ants (red), food (green), nest (blue), pheromone (yellowish). When an ant is carrying food, its color changes from red to yellow.

This week's homework uses the Kedama particle system, as discussed in class this week. Here is some documentation for an older version of Kedama. Our version is slightly different, but the general principles are the same.

Assignment

1. Run the simulation with the default controller. How much food was collected in 50 s (i.e, what was the final 'score')?

2. Rerun the simulation and answer the following questions:
a) approximately what fraction of the ants are engaged in carrying food at any given time?
b) how will a communciation system be able to improve foraging efficiency?
c) does the pattern of pheromone deposition currently provide a useful guidance cue for other ants?
d) how could the 'sender' change its own pattern of pheromone deposition to make it more useful?
e) how should the 'receiver' respond when it encounters a patch of pheromone?

Now develop your own pheromone communication systems that improves the overall foraging effiency. It should be possible to at least double the amount of food collected in 50 s. You'll need to consider the sender's pheromone deposition strategy, the receiver's pheromone response strategy, and the physical properties of the pheromone (evaporation rate). You are free to make the controller as complex as necessary to solve the task, subject to the following guidelines:

• This week your controller will not use individual sensor values; instead you have been provided with two functions called 'nestTaxis' and 'pheromoneTaxis' that will move the ant toward the nest, or uphill along the pheromone gradient, respectively.
• You ARE allowed to modify these scripts if desired, but this is probably not necessary.
• Your ant CANNOT move forward more than 1 pixel on a single time step
• The amount of pheromone that an ant deposits can be variable. It can depend on the distance of the ant from the nest (SQUEAK: "ant's distance to ant's nest") and/or can be a function of time.

3. For your final controller design, record the final score obtained on each of 3 CONSECUTIVE simulation runs (50 s):

4. What were the most challenging aspects of developing an efficient communication system?