Matrix estimation from traffic counts: integrating the proportional path averages algorithm into traffic simulation software

Abstract

Traffic simulation software packages that attempt to match assigned volumes with traffic counts rely on tried and tested demand matrix adjustment techniques. Several different methods are commonly used, most of which successively adjust the input demand matrices within iterative traffic assignment procedures. This paper follows on from a previous 2021 SATC paper wherein the author described an alternative approach based on proportional path averages, illustrated by two practical case studies. Since the algorithm could be applied independently of the assignment technique, it was implemented previously using an Excel spreadsheet containing a set of VBA macros. The algorithm requires as input only three data sets: a demand prior matrix, a table of link and/or turn traffic counts, and the assignment volumes along all Origin-Destination paths. This paper describes the direct integration of the algorithm into Emme/4 using the built-in Python scripting tools that access the Emme/4 Application Programming Interface. Three case studies illustrate the performance of the algorithm versus the standard Emme/4 demand adjustment module. The paper concludes with summary results illustrating proof of concept. Three recommendations identifying further research are also included.