Holonification model for a multilevel agent-based system

Organizational models and holonic multiagent systems are growing as a powerful tool for modeling and developing a large-scale complex system. The main issue in deploying holonic multiagent systems is the building of the holonic model called holarchy. This paper presents a novel density approach to cluster and hierarchize population in order to build the initial holarchy. The proposal extends Density-Based Spatial Clustering of Applications with Noise (DBSCAN) algorithm. Moreover, multilevel indicators based on standard deviation are proposed in order to evaluate the consistency of the holonification process. The proposed model is tested in a road traffic modeling in order to build the initial holarchy. The paper presents also the main research direction towards the control of internal and external stimuli of traffic over time.