The traffic assignment process is reported of the Delaware valley regional planning commission (DVRPC). the concepts of traffic assignment used are described and brief explanations given of the computer process. The DVRPC assignment is an all or nothing assignment which utilizes travel cost as the tree trace variable. Four sensitivity tests were made to evaluate the inputs to the traffic assignment package and the results of each test are reported. Tests 1 and 2 were unrestrained assignments. These tests were conducted to determine if turnpike tolls and accident costs should be included as components of total travel costs. Tests 3 and 4 are restrained assignments which were made to find out the advantage of calculating the 2kd value (peak-hour fraction) for each input to the capacity restraint program on the basis of generalized route and area types. The district assignment calibration indicates that the district simulated volumes are sufficiently accurate to serve as a guide in establishing the overall design requirements for freeways and high-type arterials. A zone assignment, using tests 1 and 3, was made for the portion of mercer county within the cordon area for the purpose of determining the degree of improvement in output that it has over district assignment. The 'real error' concept is introduced as a method which may be used in reporting the accuracy of the traffic assignment outputs. This method seems to be superior to the usual RMS measure used in statistics, since it takes into consideration the capacity of highways to accommodate a wide range of traffic volumes. /author/.
Abstract