[sumo-user] AW: SUMO Question

[<Pet...@dl...>]

Dear collegues,

I think the question is of general interest, so I try to add
a few remarks to Daniel's answers. Unfortunately, it's a not
a short story.=20

>
>  > If I have a network description file that describes each=20
>  edge by its free flow speed, maximum capacity, jam density.
>  >=20

First the bad news: Mikrosimulations typically do not have something=20
like a maximum capacity which can be set as an external parameter.=20

For the other two variables free flow speed and jam density, there is a=20
correspondence between maximum speed (more precisely: free flow speed is
the mean value of the desired maximum speed distribution that can be =
used=20
in most microsimulation tools -- any car might have ist onw maximum =
speed)=20
of the vehicles.=20

The jam density is simply the inverse of the generalized car length: in =
SUMO,
a car model is used where the car length, like the maximum speed, is a =
feature=20
of any individual vehicle (think of trucks and passenger cars). =
Additionally,
the car-length includes the additional space that vehicles use when =
standing
behind each other in a jam -- this is what might be called generalized =
car length.

That leaves me with the most complicated variable, the troughput or =
capacity.=20
Right now it is clear, that the macroscopic properties somehow follow =
from
the microscopic features of the vehicles. Therefore, these numbers are =
not
directly input into the simulation, they follow from the simulation. =
Nevertheless,
the most important microscopic variable that determines the maximum flow =
is
the preferred headway of the vehicles. This is what we usually call (a =
bit=20
unfortunately) reaction time, for history reasons. In principle, this =
preferred
headway can be distributed, too, but so far in SUMO it is just one =
number.=20

When regarding single lane traffic with one car-type, and without noise =
(SUMO
is a stochastic simulation, where the stochasticity can be set to zero), =
then
the maximum flow qCap reachable by SUMO is just given by:

qCap =3D 1/(tTau + lCar/vMax)

where tTau is the "reaction time", lCar is the length of a car and vMax =
is the
maximum (free flow) speed. This is the upper limit, the stochasticity =
makes qCap
smaller, because the stochastic headway becomes larger than tTau. For=20
multi-lane traffic with distributed vMax and lCar, this becomes even =
more=20
complicated, to my knowledge, there are not closed analytically formulas =

available that relate qCap with the parameters of the vehicles (for this =
simulation
model). Simulation will tell, and this demonstrates, that capacity in =
reality=20
is never a fixed number, because what I described here pretty much =
applies to=20
reality itself, not just to simulation.

To be practical: the most two most important variables that determine =
qCap in
SUMO are tTau and the parameter of stochasticity, eps. Some =
experimentation is
needed to find out which set of parameters gives the right capacity.=20

Hope that answers all the questions. If not, do not hesitate to ask.

Best regards,

   Peter



Dr. Peter Wagner
Institute of Transport Research
(Institut fuer Verkehrsforschung)
German Aerospace Center (DLR),=20
12489 Berlin-Adlershof, Germany
Phone: +49/30/67055-237
Fax:   +49/30/67055-202
eMail: pet...@dl...=20
 =20

>  -----Urspr=FCngliche Nachricht-----
>  Von: Krajzewicz, Daniel=20
>  Gesendet: Mittwoch, 1. Dezember 2004 12:41
>  An: wal...@eg...; sum...@li...
>  Cc: Wagner, Peter
>  Betreff: AW: SUMO Question
> =20
>  Dear Walaa,=20
> =20
>  > -----Urspr=FCngliche Nachricht-----
>  > Von: wal...@eg... [mailto:wal...@eg...]
>  > Gesendet: Mittwoch, 1. Dezember 2004 09:42
>  > An: Krajzewicz, Daniel
>  > Betreff: SUMO Question
>  >=20
>  > Dear Mr. Daniel,
>  >=20
> =20
>  Please call me just "Daniel" - that's my first name...
> =20
>  >   I'm using SUMO in my reasearch simulations in Computer=20
>  Science. But=20
>  > I have some few Questions:
>  >=20
>  > If I have a network description file that describes each=20
>  edge by its=20
>  > free flow speed, maximum capacity, jam density.
>  >=20
>  > How can I set such properties to an edge in the edges.xml=20
>  file i.e. :
>  >=20
>  > 1- Can the speed attribute in the edges.xml (which is the max=20
>  > "allowed" speed) represent the free flow speed? If not how can I=20
>  > introduce the free flow speed to SUMO?
>  >=20
> =20
>  Well, I suppose so. As you know each edge has a maximum=20
>  allowed speed. This speed can of course be only driven at=20
>  the edge if it's empty enough. So one could say that the=20
>  maximum speed is the free flow speed. On the other hand,=20
>  vehicles may drive slower than the speed allowed on the edge=20
>  - either because of the vehicle's maximum speed or because=20
>  the vehicle "dawdles"...
> =20
>  > 2- How can I express the edge's max capacity and jam=20
>  density in SUMO?
>  >=20
> =20
>  I don't know where you need it. The maximum capacity of an=20
>  edge is somehow proportional to the number of lanes=20
>  (~2400vehicle/h is a highway lane's maximum capacity). And I=20
>  don't know the term "jam density" or at least the values for=20
>  it. Maybe anyone else from the sumo-users (mailinglist) will=20
>  be able to answer this.
> =20
>  > Thanks in advance,
>  > Walaa.
> =20
>  Bye,
>  Daniel
> =20
>  >=20
>  >=20
>  >=20
>  >=20
> =20