by
Andrew Cotugno, Dick Walker and Nina Kramer
Metropolitan Service District
2000 S.W. First Ave, Portland OR, 97201-5398
The Metropolitan Service District (Metro) recently assisted a consultant hired to review the functional classification system and state route designation for the roadways of this region. By taking advantage of the versatility of the auto assignment module within EMME/2, a rather unique approach was developed to determine the type of trips that use a particular facility. This article deals with the concepts of that approach.
Introduction
The question to be asked was simple: "Who are the primary users of a roadway - regional or local trips?" Metro's role in the study was to generate some technical data to help formulate and support a position. The tasks were twofold: 1) develop two trip tables - local vs. regional; and 2) develop an assignment procedure in which the two trip types can be kept separate but yet collectively result in a reasonable loading.
Regional vs. Local Trips
It was decided that the p.m. peak-hour trips would be used as the universe for the study. Hence, all trip table stratifications and assignments would be made for that time period.
It was further determined that "regional" trips would be defined as those trips having a length greater than the average trip length for the p.m. peak-hour. Using module 3.16, Plot Matrix Histograms, the average trip length was found to be 6.4 miles. The local trips were defined to be those whose length were less than 6.4 miles.
The distance matrix was used as the constraint matrix in module 3.21, Matrix Calculations, to stratify the total p.m. peak-hour trip table into the two components defined above. The resulting p.m. peak "regional" trip table had 89,600 interzonal trips in it; the "local" trip table had 140,600 interzonal trips.
Assignment Technique
For this study, it was decided that the best approach would be to use the local trips as background traffic for a full equilibrium assignment of the regional trip table. This was accomplished in the manner described below.
First, a partial trip table (0.75 * total p.m. peak trips) was assigned to the network using the full equilibrium algorithm. Module 2.41, Network Calculations, was used to transfer the resulting link volumes to a user data item. Referencing this field as additional volume in the assignment procedure, an all-or-nothing (0 iteration) assignment was made with the "local" trip table. Hence, based on travel times derived from a partially congested network, the local trip link volumes were established.
It was hypothesized that the all-or-nothing assignment for the local trips would be reasonable because the alternative path choices for short trips are not as numerous or as diverse as those for longer trips. Furthermore, a full equilibrium assignment did not seem reasonable because there were not enough trips to adequately reflect peak-like congestion levels.
The local trip link volumes were then moved to a user data item. Referencing these local volumes as additional volumes, a full equilibrium assignment was performed on the "regional" trip table. This process allowed the local trip link volumes to remain fixed while the regional trips sought alternative paths to achieve network equilibrium.
Assignment Validation
The total auto volumes of this assignment were compared to those obtained from a previously calibrated full p.m. peak trip table equilibrium assignment. Using module 6.13, Compare Auto Volumes, bandwidth plots were made to identify the differences. The assignment compared quite favorably. Very few links varied by more than 100 vehicles; and for those which did, the results were within 10 percent.
A plot was made of the auto volumes only (i.e. the regional trips) using module 6.12, Plot Auto Volumes. The flows were reviewed for reasonableness and continuity.
Analysis of Results
The Network Calculator (module 2.41) was used to determine the ratio
between regional trips and the total auto volume for each link. After
first initializing a user field, this was accomplished by using the
expression "volau/(volau+volad)", where volau represents the regional
trips and volad represents the local trips. The calculation was only
performed for those links where volau did not equal zero. This was
done to guard against division by zero. If volau was zero, the result
would also assume zero due to the initialization of the user field.
A bandwidth plot of the ratio was done using module 2.13, Plot Base Network. The plot graphically illustrated those facilities carrying the higher percentage of regional trips.
Summary
The strategy presented above is only one of many valid techniques that could be used to gain a better understanding of regional vs. local trip-making patterns. However, it does serve to illustrate the flexibility of the tool and its use in the solution methodology. Through the use of the Network Calculator, partitioned trip tables in the assignment process, and judicious use of bandwidth plots, the travel patterns within a region can be analyzed in a very discerning manner.