Hi Jane,
There are quite a few
ways to do this. What we have been doing while working on demand
calibration methods is to have a wrap-like functions that do all
the estimations, save the OD file, and then run all SUMO procedures
from that point on.
We have tested implementations
for both Matlab and R and it is quite easy to do in any
programming language. As you can understand this might or might
not run traci. In our case, we do not use it.
The "secret" to doing
everything (that worked for us) is the `system` command. With
this we use the SUMO tools and get we need. You can have a look
at the code below, which provides the logic and is pretty much
the core of running sumo this way:
High level description:
- First we are
creating the OD file
- then we do time formatting
- set the seed
- convert OD2TRIPS
- RUN trip based sumo
- Convert output to
csv (to make it easier to read in matlab)
- Read output in
Matlab
Matlab important pieces
of code (just for guidance, not guaranteed to work as is):
% CREATE the file
format to pass from a matrix to an OD
TheBeginText= strcat("$OR;D2 \n* From-Time To-Time
\n",starttime," ", endtime,"\n* Factor \n1.00\n");
Expr = zeros(size(FullOD,1)*size(FullOD,1),3);
Expr = string(Expr);
Counter=1;
for i = 1:size(FullOD,1)
for j = 1:size(FullOD,2)
Expr(Counter,1) = Zones_masks(i,2);
Expr(Counter,2) = Zones_masks(j,2);
Expr(Counter,3) = string(FullOD(i,j));
Counter=Counter+1;
end
end
% SAVE the OD File in the temp location (each file gets a k for
the par loop)
ma=strcat(pathtoCaseOutput,'/',num2str(k),'myODUpdated.txt');
fileID = fopen(ma, 'w');
fprintf(fileID,TheBeginText);
fprintf(fileID,'%s %s %s \n', [Expr(:,1),Expr(:,2),Expr(:,3)]');
fclose(fileID);
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
TIME FORMATTING %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
% Start Simulation Times
integ=floor(str2double(starttime));
fract=100*(str2double(starttime)-integ);
beginSimTime = integ*60*60 + fract*60;
% End Simulation Times
integ=floor(str2double(endtime));
fract=100*(str2double(endtime)-integ);
endSimTime = integ*60*60 + fract*60;
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
SEED SETTING %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
seedNN = randi(100000, 1);
%%%%%%%%%%%%%%%% MAC /
LINUX OD2TRIPS %%%%%%%%%%%%%%%%
oD2TRIPS = strcat(pathtoSUMOBIN,"od2trips --no-step-log
--output-prefix ",num2str(k),...
" --spread.uniform ",...
" --taz-files ", pathToNetwork_a,tazname, " -d ", ...
pathtoCaseOutput,"/",num2str(k),"myODUpdated.txt -o ",
...
pathtoCaseOutput,"/","upd_ODTrips.trips.xml",...
" --seed ", num2str(seedNN));
[~,~] = system(oD2TRIPS); % [~,~] = suppresses any console
outcome. Not recommended when running first time
%%%%%%%%%%%%%%%% TRIPBASED RUN
%%%%%%%%%%%%%%%%
sumoRUN=strcat(pathtoSUMOBIN,"sumo
--mesosim --no-step-log --output-prefix ",num2str(k)," -n ",...
pathToNetwork_a,mesoNet, ...
" -b ",num2str(beginSimTime),...
" -e ",num2str(endSimTime+500),...
" --load-state
",pathtoCaseOutput,"/",num2str(iter-1),"state.xml" , ...
" --save-state.times ", num2str(endSimTime), ...
" --save-state.files ",
pathtoCaseOutput,"/",num2str(iter),"state.xml", ...
" -r ", pathtoCaseOutput,"/", num2str(k),
"upd_ODTrips.trips.xml",...
" --additional-files ", pathtoCaseOutput,"/",
additionalName,...
" --xml-validation never",...
" --device.rerouting.probability
",string(rerouteProb),...
" --seed ", num2str(seedNN));
[~,~] = system(sumoRUN); % [~,~] = suppresses any
console outcome. Not recommended when running first time
%%%%%%%%%%%%%%%%
Read Output %%%%%%%%%%%%%%%%
Data2csv =
strcat("python ", pathToSUMOtools,"/tools/xml/xml2csv.py",...
" ", pathtoCaseOutput,'/' ,loopDataName , " -o ",
pathtoCaseOutput,'/' , loopDataName,".csv",...
" -x ", pathToSUMOtools,'data/xsd/det_e1meso_file.xsd');
system(Data2csv);
newStr =
strrep(pathtoCaseOutput,"\","");
mDir = char(newStr);
KKname=strcat(mDir,'/',loopDataName,".csv");
simulated_tripsInTable = readtable(KKname);
simulated_tripsInTable(simulated_tripsInTable.interval_end>endSimTime+1,:)
=[];
% The final step is to aggregate the counts per link (so that we
do can
% estimate per link flow and not per lane).
mm=simulated_tripsInTable.interval_id;
newStr = split(mm,'_');
myEdges = newStr(:,2);
simulated_tripsInTable.EdgeID = myEdges;
G = findgroups(simulated_tripsInTable.EdgeID);
AllPeriodEdgesFlows = unique(myEdges);
AllPeriodEdgesFlows(:,2) =
num2cell(splitapply(@sum,simulated_tripsInTable.interval_entered,G));
I hope this helps!
Best,
Manos
--
Chaniotakis Manos (Emmanouil)
Technical University of Munich
Department of Civil, Geo and Environmental Engineering
Chair of Transportation Systems Engineering
e:mail: m.chaniotakis@xxxxxx
website: https://www.tse.bgu.tum.de/en/staff/mchaniotakis/
Tel: +49 89 28910461
On 29/10/2018 16:07, Jane Cheung wrote:
![]()
Hi guys, there is
one problem that
I would like to listen to your
suggestion. Appreciated in advance!
The thing is that how to simulate different scenarios,
one by one. It does not need to click the renewed
configuration. In other would, the route files or other files
could be generated automatically and the new .sumocfg could be
generated.
There is an original configuration in my file and run it,
the interested variables (like the mean travel time in one
edge) in the whole solution duration could be got .Then,
other options would be done (by some TRACI or some other
methods) to modify the route files or related files and the
renewed one could be generated (like the density is increasing
or the width of the edges is becoming larger or some others)
by TRACI , then, call the SUMO command line
again, run the .sumocfg again. Repeat the simulation
again and again to the get the changing trend of
variable you are interested in.
Simplely, could some options be called to renew the
route.xml automatically?
Thanks in advance!
Best regards!
Jane
_______________________________________________
sumo-user mailing list
sumo-user@xxxxxxxxxxx
To change your delivery options, retrieve your password, or unsubscribe from this list, visit
https://www.eclipse.org/mailman/listinfo/sumo-user
