A method is described to connect microscopic, mesoscopic and cellular automaton based traffic models in order to allow the model's concurrent simulation. The traffic models are characterized by different levels of modelling detail and distinct simulation platforms. The approach includes the assignment of appropriate simulation tasks for the model components as elements of a complex hybrid model. Furthermore, implementation related aspects of network parts and the concurrent simulation of the hybrid model by means of separate simulation kernels are discussed. New approaches of simulation technology as High Level Architecture (HLA) were applied and adapted with regard to design of interfaces and synchronization management. A realistic model of an urban traffic network was developed as a reference and it demonstrated how to build up and transform each of the model types. To solve this task, the applicability of Geographic Information Systems was examined and in conclusion used for the purpose of model design.
Abstract