Information Technology Reference
In-Depth Information
Evolving: Computer Science Models.
Previous research in computer sci-
ence has also investigated the replication of emerging phenomena observed in
social reality, while keeping low agents simple. The most famous ones being Sug-
arscape [5] and New Ties [6]. Sugarscape [5] introduces a very simple grid-like
environment and simple agents with simple rules. But, in adding slight varia-
tions on top of this environment and agents (like reproduction, fight and trade),
Epstein and Axtell were able to replicate a wide range of observed social phe-
nomena (e.g. genetic selection, global wars, real-world market behaviors). New
Ties [6] aims at building an artificial society with socially high mental and lin-
guistic capabilities, but with simple agents and simple environment. The goal
of New Ties is to build autonomous entities, living and adapting to the world
due to learning. Moreover, these agents are capable to produce their own culture
and share it with other agents. The research objectives of New Ties aim at a
better understanding of culture formation and linguistics. In this article, our
research differs from them in the sense that we aim at building societies capable
of dynamically expanding themselves with explicit institutions, but still in their
bottom-up approach.
4 A Model for Evolving Societies
Our model consists of two main features: the first concerns environmental aspects
(e.g. food, reproduction) and the second is focused on social aspects (human
interactions). Environmental aspects represent food collection and reproduction.
Well fed human agents reproduce. Thus, the society grows until the food input
matches the survival costs of humans. If the food input is lowered, agents starve
to death and the society becomes smaller.
Social aspects concern human agents social interaction. In particular, it in-
troduces social issues caused by population scale changes. When the population
is bigger than a family, family-unrelated agents are likely to fight with each
other when the food is sparse. In this case, a police organization is created
when the population feels globally unsafe. This leads the society to the chiefdom
stage. Similarly, when the population increases, the loyalty towards the leader
decreases. When the loyalty is too low, a
CH
organization is created to keep
high loyalty, leading the society to the state stage. As the chiefdom stage arises
before the state stage, state is also chiefdoms. In the following, the model is pre-
sented in more detail using the SugarScape [5] formalism. This formalism splits
the model in three sections: environment, agents and model rules.
4.1 Environment
2
space for agents
The environment is similar to the Netlogo space: a bounded
R
2
grid). Each patch
has 3 states: (
sterile,harvested,full
). A
sterile
patch never changes state.
A
full
one can be harvested by human agents, thus becoming
harvested
.A
harvested
cell can become
full
each round (
p
=0
.
005). In our experiments,
patches in a radius of 8 around a village are initially
full
.Othersare
sterile
.
on top of which food is represented by NetLogo patches (a
N
