This page is no longer up-to-date. Please visit amybia.loria.fr

AMYBIA
Aggregating MYriads of Biologically-Inspired Agents
AMYBIA screen

Simulations

Despite the simplicity of its local update rules, the global behaviour of this system is hard to predict if the number of interacting agents is large. In other words, it is a complex system. The study of complex systems generally requires large amount of data; this is particularly important when the system displays phase transitions, i.e., a brutal change of behaviour triggered by a small change in the value of one parameter. A first issue is thus related to generating and analysing such large amounts of data. A second major issue concerning these simulations is the evaluation of the robustness of the model : it is important to make simulations with different conditions? What happens to the simulations if the update policy is changed? What happens to the simulation if the topology of the grid is modified? etc. Once again, addressing this issue is a critical step on the path to utilising the model properties in the framework of distributed computation devices.

Last Update : Nov 2008