2012 Project 4
From Autonomous Multirobot Systems
Contents |
MASON Predator Prey with Communication
Overview
In this project you will extend your predator (or prey) code to use communication. You should assess your new system in terms of the performance gain provided by the additional communication. You should make (at least) 10 runs of your old system and record the performance, then 10 runs of your new system and report the mean and standard deviation of the performance of each.
Other variations are OK, but please confirm with me first.
What to Do
You should have your predator prey code working with your algorithm. You will now add communication. I want you to think about how you can gain the most performance from the least amount of communication. I want you also to assess or estimate the amount of information exchanged at each step of the simulation.
Make sure your old code now works in this new context. Then add communication. Then assess both systems as described above:
- Assess the approach in terms of performance-based experiments (measure the performance) -- 10 runs on each system, Mean and standard deviation.
- Create a one to two page writeup describing your system.
Expectations, What not To communicate
Your new system should perform better than your previous system as measured by total time until all prey are captured. (If you are coding prey, your prey should live longer)
Hmm, for now I can't think of anything that wouldn't be OK to communicate :-), but predators cannot peek at prey data and vice versa.
What to Turn In
You should submit a screen shot of the predator/prey system running on your computer:
- XXcontroller.java the code for your controller
- writeup.pdf your writeup
