Calibrating Large Scale Microscopic Simulation: Recent Pulication

This recent paper might be a useful reference 

Large-Scale Microscopic Simulation: Toward an Increased Resolution of Transportation Models

J. Transp. Engrg., Volume 134, Issue 7, pp. 273-281 (July 2008)

Mark C. Smith,1 P.E.; Adel W. Sadek,2 A.M.ASCE; and Shan Huang3
1Research Assistant, School of Engineering, Civil and Environmental Engineering Program, Univ. of Vermont, Burlington, VT 05405. E-mail: msmith@uvm.edu
2Associate Professor, School of Engineering, Civil and Environmental Engineering Program, Univ. of Vermont, Burlington, VT 05405 (corresponding author). E-mail: asadek@cems.uvm.edu
3Research Assistant, School of Engineering, Civil and Environmental Engineering Program, Univ. of Vermont, Burlington, VT 05405. E-mail: shuang@uvm.edu

(Accepted 26 December 2007)

Recent years have witnessed an interest in increasing the resolution of transportation models to allow them to better address emerging transportation issues, such as concerns about emissions, air quality, and emergency preparedness. Microscopic simulation models, especially those that combine routing logic with accurate modeling of traffic flow dynamics, have vast untapped potential for modeling large-scale transportation networks and for achieving the resolution level required for addressing such emerging issues. Building and calibrating such models, however, is quite challenging and is currently not widely understood. This paper describes the process of developing and calibrating a large-scale microscopic model of Chittenden County, Vt., an area of about 1,400 km2 (540 mi2). The model was developed primarily using readily available data to which most metropolitan planning organizations are expected to have access. Following a brief description of how the network was coded and the traffic demand specified, the paper discusses the process of error checking and model calibration. Preliminary calibration results are encouraging, given the complexity of the model. The paper includes a summary of the main lessons learned and modeling pitfalls that should be avoided.

©2008 ASCE