Box-Jenkins time-series analyses were conducted on 20-, 40-, and 60-s occupancy and volume data collected during a morning rush period at two freeway locations--one on the local lanes and the other on the express lanes on the Dan Ryan Expresseay in Chicago, Illinois. The local section experienced heavy truck traffic and was connected to an entrance ramp. The express section, on the other hand, was of the straight-pipe type that had no truck traffic. Several autoregressive integrated-moving-average (ARIMA) models were evaluated, and the ARIMA (0, 1, 1) model was found to be the most statistically significant for all forecasting intervals for both volume and occupancy. A comparison between the ARIMA (0, 1, 1) model and the Illinois Department of Transportation Traffic Systems Center ARIMA (0, 1, 0) model showed significant improvement in forecasting for the former. The 60-s forecasting interval was found to be the most-effective interval. When compared with occupancy forecasting, volume forecasting was found to be less variable than occupancy, as expressed by the ratio of the residual sum of squares to the mean of the observations. By the same token, forecasts of volumes and occupancies on the express lanes were found to be less variables than those on the local lanes. (Author)

Media Info

  • Features: Figures; References;
  • Pagination: p. 47-49
  • Monograph Title: Grade crossings, devices, visibility and freeway operations
  • Serial:

Subject/Index Terms

Filing Info

  • Accession Number: 00334176
  • Record Type: Publication
  • ISBN: 0309031176
  • Files: TRIS, TRB
  • Created Date: Oct 28 1981 12:00AM