I am modeling a rural highway that is predominantly a 2 lane corridor, with Passing lanes and 1 isolated (un-coordinated) actuated signal. The corridor is about 8 miles long and the goal is to determine tiome savings by adding passing opportinities in the form of longer or additional passing lanes.
While observing a recorded simulation, I was surprised to find trucks going well slower than they should have been. A little background: In California, trucks and cars w/ trailers are legally limited to a speed of 55mph. The speed limit on the road is 55, and our speed surveys indicated the majority of trucks travelling 55-60mph, with an overall average speed of 62mph. The speed survey data points were broken down to determine speed factors for driver types, based on the avergae 62 mph.
This would of course result in truck exceeding the speeds they were observed at, and so the truck profiles (Semi Truck 1, 2, TruckDB) were modified to reflect a lower top speed of 60 mph.
So at one point while observing passing behavior I noticed a TruckDB travelling at 45mph. No combination of average speed and driver behavior (speed modifier) would result in such a slow speed, so it puzzled me. Some of the Semi trucks were travelling at slower speeds than their driver type would have indicated, about 51mph and a driver type 10!! Shouldn't he have been pushing 60?
So I went to the entrance to the network at the southern end and the trucks were entering the network at the proper speeds, like clockwork: (speed factor)x(link speed) So I followed a truck. what happened is that when a truck slowed, or stopped for the traffic signal, they would not accelerate back to the previous speed of (speed factor)x(link speed), they got to the previously observed lower speeds and did not accelerate any faster - driver type did not appear to influence how fast they got to, the majority determination seemed to be based on acceleration. The 2mph acceleration trucks would get to about 45mph, and the 3mph trucks would get to a little faster. This proved consistant when changing the acceleration rates between different truck types. Experimenting with different top speeds for the trucks revealed that by adjusting the max vehicle speed upwards, the 45mph speeds adjusted upwards...trucks would accelerate to some faster speed instead of 45mph. But with higher top speeds, the trucks would enter the network at higher speeds - much faster than they would have been travelling.
My solution was to insert a "dummy" link at each end of the network at 45mph, so that vehicles enter the network at a slower speed, and allow them to accelerate to the proper speed after moving to the next higher speed link. trial and error produced results with the trucks getting up to 55-60 mph.
I believe what is happening, after reading the help file on vehicle profiles is this: according to the help file acceleration is reduced linearly based on current speed vs. maximum speed, with max accel at 0 speed and 0 accel at max speed. so it should be a straight line from 2 or 3 mph accel at 0 ft/sec to 0 mph at max speed. there must be some kind of cutoff or rounding though; my observations of the vehicle's showed the accel dropping almost instantly from 1mph to 0mph, thus "capping" the speed.
Can the developers offer any insight?

Eric Royer