La complexité dans les Simulations Multi-Agents

The algorithmic complexity of a multi-agent simulation comes from several sources : the intrisic complexity of the phenomenon to tackle, cognitive skills of agents as well as the mode- ling choices made while designing the distributed model and the underlying simulation engine (behaviour, environment, time, perception models, etc.), each one being likely to be implemented by multiple ways. It is necessary to explain each of those choices, from analysis to implementation, to optimize the progression of the resulting computer simulation, to avoid the introduction of sometimes critical biases and to obtain a simulation as simple as possible for a given phenomenon.