Treffer: Parallelizing DEVS Traffic Model for Distributed Systems.

As traffic simulations increase in complexity and scale, discrete event-based approaches demonstrate significant performance advantages, particularly when deployed in clustered and distributed computing environments. This paper introduces a parallelization strategy for the DEVS-based traffic model to improve scalability and reduce execution time. Experimental results demonstrate a substantial reduction in simulation runtime, even within a single-node configuration. To further enhance scalability, it is proposed a cluster-based architecture in which atomic models are distributed across multiple nodes, coordinated by a central controller. Our findings highlight the effectiveness of parallelization in DEVS-based traffic modeling and lay the groundwork for future research on fully distributed traffic simulation systems that show acceptable accuracy while achieving high performance at large scales. [ABSTRACT FROM AUTHOR]