Ad Hoc Distributed Dynamic Data-Driven Simulations of Surface Transportation Systems

  • Authors:
  • Michael Hunter; Hoe Kyoung Kim;Wonho Suh;Richard Fujimoto;Jason Sirichoke;Mahesh Palekar

  • Affiliations:
  • School of Civil and Environmental Engineering GeorgiaInstitute of Technology, GA, USA;School of Civil and Environmental Engineering GeorgiaInstitute of Technology, GA, USA;School of Civil and Environmental Engineering GeorgiaInstitute of Technology, GA, USA;Hewlett Packard Colorado Springs, CO, USA;Hewlett Packard Colorado Springs, CO, USA;Google Inc Mountain View, CA, USA

  • Venue:
  • Simulation
  • Year:
  • 2009

Quantified Score

Hi-index 0.00

Visualization

Abstract

An ad hoc distributed dynamic data-driven simulation is a collection of autonomous online simulations brought together to model an operational system. They offer the potential of increased accuracy, responsiveness, and robustness compared to centralized approaches. They differ from conventional distributed simulations in that they are created bottom-up rather than top-down. They combine concepts from conventional distributed simulations and replicated trials, raising new issues in data management and synchronization. In this article, the ad hoc simulation approach and an optimistic synchronization algorithm are proposed. A prototype coupling in-vehicle transportation simulation is evaluated and shown to yield results comparable to a traditional replicated experiment for the tested scenarios. Experiences applying this concept to a commercial transportation simulator in an emergency scenario are described.