Sugarscape
Pedro Andrade
Gilberto Câmara
Agent-Based Modelling
Representations
Goal
Communication
Communication
Action
Perception
Environment
source: Nigel Gilbert
The Sugarscape Model
Building artificial societies that
model certain characteristics
of real ones
Cells produce sugar: rule Gα
Each cell grows α
units of sugar by
time step up to
capacity
Sugarscape: Agents consume sugar
Agents look for
sugar to eat
Agents attributes
(age, wealth)
Sugarscape: Simple agent
Look for a random
neighbor
Moves there if it has
more sugar than the
current cell to eat.
Agents carry food with them
Agents have an initial quantity of food
They also carry food that has not been eaten
Agent survival
Agents collect sugar and consume
When they have no sugar, they die
Agent movement rule
Look out as far as vision enables – find the closest
empty place with the most sugar
Move there and get all the sugar
Simple Agent

Maximum sugar between 0 and 4

Each cell produces one unit of sugar by time
step

Each agent starts with initial sugar of 10 and
needs one sugar per time step

Look for a random neighbor and moves
whenever it has more sugar than the current
cell to eat.

getSugar(“default.sugar”) -- logo package

Attribute “maxSugar”
More complex Agent

Look for all neighbors and moves to the empty one
that has more sugar.

Initial wealth (sugar) between 5 and 25

Metabolism between 1 and 4 sugar per each step

Age starting with zero and increasing by one per
time step

Die by age (maximum age between 60 and 100)

Die by absence of wealth

Whenever it dies, creates another agent and put it
in a random cell
Changes

What happens with the average wealth and
average age when the number of agents grow?