Ants - Scalable Game Design wiki by yurtgc548


									Computational Representation
     of Ant Foraging
         Clayton Lewis
         June 26, 2010
Find and Download the “Ants”
• from the scalable game design home
  – click “Summer Institute 2010”
  – then “STEM Simulations”
  – then “Ants”
  – then “complete” to download Agentsheets
• Run the complete simulation, and see what
  features of the behavior of the simulated ants
  you can identify.
• At this stage, don’t examine the rules, or
  modify simulation properties.
• Run the simulation fairly slowly at first so as
  to be able follow the behavior of individual
• Think about the brown markings as trails, not
• How do these behaviors of the
  simulated ants compare to those of real
Thought Questions about Real
• How do they find their way back to the
• Why could it be useful for the trail
  pheromone to evaporate?
       Exploring Variations
• The simulation property “ticks”
  measures how quickly the ants have
  exhausted the food.
• The simulation property “turning”
  controls how often ants turn when
• Let’s use these properties to explore
  what value of turning works best
          Exploring Turning
• We’ll create a simple food layout
• Use the arrow tool and move the nest to the
  middle of the worksheet
• Erase the food
• Place 4 lollipops a little way NE,SE,SW, and
  NW of the nest
• Erase the ants and put 4 in next to the nest,
• SAVE the worksheet
   Exploring Turning (Cont.)
• We’ll assign values of “turning” by
  counting off: 10,50,90
• Reset your simulation
• Set the value of turning to your value in
  the Simulation Properties Window
• Run the simulation
• Report the value of ticks when all the
  food is gone
What is the best turning
    Exploring Turning (Cont.)
• Let’s change food layout
• Use the arrow tool and move the food near
  the corners of the worksheet

• Check that your simulation properties window
  shows your value of turning, and ticks 0
• Run the simulation again with your value of
What is the best turning
        More Explorations
What is the effect of pheromone life?
 Does it depend on the amount of food at
 a food site?
Does this effect depend on number of
Does the optimal turning value depend on
 the number of foragers?
Actual Ant Turning is Density
 Dependent (Gordon, 1999)
        About the Program
• The program uses several
  computational thinking patterns:
  – collision (ant and food)
  – diffusion (distance to nest)
  – hillclimbing (to find nest)
  – polling (to control ticks)
• These and many others are described
  on the Wiki
     Extending/Changing the
• What if there is no pheromone trail?
  – (Try it… first find the rule that places the trail… a
    little tricky!)
• The ants don’t follow the trail perfectly… can
  you (or your students) improve this?
• Some real ants don’t begin to search until
  they have moved some distance straight from
  the nest…(why?) … could you add this
          Ant Tournament
• Students can be challenged to produce
  improved ants that forage more quickly
• …and to research how their
  improvements relate to real ant foraging
            Final Words
• Real ant researchers use simulations to
  understand ant behavior…
• … and some computer scientists study
  ants to understand new ways of
  searching for information.
• More ant materials are on the Wiki

To top