Automated Optimal Balancing of Traffic Volume Data for Large Access-Controlled Highway Networks and Freeway-to-Freeway Interchanges

Traffic analysts and modelers frequently need to combine street and highway volume data from multiple days and sources in order to assemble a complete data set for a study area. Pro-rata methods have traditionally been used to “balance” inconsistent data sets for limited-access facilities, but the pro-rata procedure has several limitations and cannot readily be applied to complicated roadway geometries such as freeway-to-freeway interchanges. Mathematical optimization of the data set using algorithms such as the GRG2 (Generalized Reduced Gradient) solver can address many of these situations and allows rapid balancing of large roadway networks for multiple time periods. Optimization also allows a second (perhaps less reliable) set of mainline observations to be considered in the balancing process. The authors selected minimization of the sum of the GEH (Geoffrey E. Havers formula) for each observation as the objective function and implemented the procedure in a spreadsheet (it could also be implemented as a database-driven application). Optimization produced more favorable GEH values than pro-rata methods, indicating better fit to the original data set. Optimization results combining GRG2 and GEH appear to be reasonably stable and tolerant of fluctuations in the input data values attributable to typical traffic counting devices. Additional research is necessary to quantify the effects of severe outliers on the optimization results and determine which combination of objective function and search algorithm is “best” for various types of networks.

Language

  • English

Media Info

  • Media Type: Digital/other
  • Features: Figures; Maps; References; Tables;
  • Pagination: 12p
  • Monograph Title: TRB 93rd Annual Meeting Compendium of Papers

Subject/Index Terms

Filing Info

  • Accession Number: 01520088
  • Record Type: Publication
  • Report/Paper Numbers: 14-3565
  • Files: PRP, TRIS, TRB, ATRI
  • Created Date: Mar 26 2014 10:13AM