Information Technology Reference
In-Depth Information
Enhancing Cellular Spaces by Multilayered
Multi Agent Situated Systems
Stefania Bandini, Sara Manzoni, and Carla Simone
Department of Informatics, Systems and Communication
University of Milano-Bicocca
Via Bicocca degli Arcimboldi 8 20126 Milan - Italy
tel +39 02 64487835 fax + 39 02 64487839
{
bandini, manzoni, simone
}
@disco.unimib.it
Abstract.
This paper presents the formal description of the Multilay-
ered Multi-Agent Situated System (MMASS)
1
which can be seen as a
generalization of cellular spaces since it relaxes some constraints on uni-
formity, locality and closure. MMASS allows the description, represen-
tation and simulation of complex systems that explicitly require to con-
sider spatial features. Different forms of interaction are possible within
a MMASS: synchronous reaction between spatially adjacent agents and
asynchronous and at-a-distance interaction through a field emission-
propagation-perception mechanism.
1 Introduction
Cellular Automata (CA) can be seen as a kind of Multi Agent System (MAS),
where spatial structure of agent environment is explicit and structured in a grid,
agents are immobile, homogeneous and dense (all the cells of the CA are identical
and include agent representation) and their behavior is synchronous. This is not
the general case in MAS-based models, since heterogeneous and asynchronous
agents might live in the same, possibly not structured, environment. Thus, as
clearly stated in [1], “CA can be considered either as 'degenerate' MAS in which
agents have become fixed or, more positively, as good environmental models”. A
MAS consists of a number of agents that are defined in terms of their behaviors
and characteristic parameters and are located in an environment that makes their
interactions possible. The behavior of a MAS is defined as the global effects of
local interactions among autonomous agents populating the environment.
Accordingly, Multi Agent Based Simulation (MABS [2]) is based on the idea
that it is possible to represent the global behavior of a dynamic system as the
result of interactions occurring among an assembly of agents with their own op-
erational autonomy. MAS have been used to simulate artificial worlds [3] as well
as natural or social phenomena [4]. In MABS, agents might represent animals
1
The work presented in this paper has been partially funded by the Italian Ministry of
University and Research within the project 'Cofinanziamento Programmi di Ricerca
di Interesse Nazionale'
