Towards a Multilevel Simulation Approach Based on Holonic Multiagent Systems

Simulation is an appropriate approach for study-ing complex systems that are inacessible through di-rect observations and measurements. In a simula-tion involving a great number of interacting entities, it is difficult to create a reliable and tractable ab-straction of the real reference system. One of the involved problems is amount of computational re-sources required to handle microscopic simulation of large number of entities. One solution is to use macroscopic models. However, this type of models may be at hand unavailable or not reliable, or it doesn’t allow observations of individual behaviours. In this paper a multilevel simulation model is pro-posed to dynamically adapt the level of simulated be-haviours while being as faithful as possible to the reference model. Our approach is based on Holonic Multi-Agent Systems and provides a generic schedul-ing model for multilevel simulations.