Then I would use libsumo:
- load a weight file corresponding to a jammed network
- run lots of simulation.findRoute queries (and measure time)
- initialize your own algorithm graph from the sumo data structures
- run the same queries against your own routing algorithm (and measure time)



Am Do., 27. Feb. 2020 um 03:23 Uhr schrieb Raheleh Zarei <
[email protected]>:

> The basic comparison I want to do is how fast they find shortest path when
> there are many s-t queries and do rerouting (finding a new route) when
> there are traffic jam in some edges.
>
> On Tue, Feb 25, 2020 at 1:43 AM Jakob Erdmann <[email protected]>
> wrote:
>
>> Given the distance/travelTime data for each edge, the shortest/fastest
>> path in a typical network  is often unique and thus all algorithms reach
>> the same result (manhatten grid networks are an obvious exception to this).
>> Therefore it does not matter which algorithm you use as long as they use
>> the same inputs.
>> Thus, I was trying to understand what difference you would expect from
>> using non-sumo algorithms.
>>
>> Am Di., 25. Feb. 2020 um 05:02 Uhr schrieb Raheleh Zarei <
>> [email protected]>:
>>
>>> I am not sure if I understand the first paragraph of your email. The
>>> algorithms are shortest/fastest path finders(I call them non-sumo
>>> algorithms). Actually the thing I want to do is as follow:
>>> -Import map of a city as road network
>>> -Make some source and target points based on the map
>>> -Run random number of source-target queries between these source and
>>> target points and for every query (or in a time interval for example every
>>> 20 second) get the current situation of the network (like the travel time
>>> of edges because it might have changed due to heavy traffic in some edges)
>>> and find the shortest path/fastest using non-sumo algorithms.
>>> -calculate the path found (time/distance) by non-sumo algorithms for
>>> each source-target query and also average of time-distance for the whole
>>> s-t queries during simulation.
>>>
>>> Thanks for your help,
>>> Raheleh
>>>
>>>
>>>
>>> On Mon, Feb 24, 2020 at 2:11 AM Jakob Erdmann <[email protected]>
>>> wrote:
>>>
>>>> It depends on what these other algorithms are. If they are also
>>>> shortest/fastest path finders that it would be sufficient to alter the
>>>> weights/travel times assigned to the network edges (via TraCI or xml
>>>> inputs) and keep using the sumo algorithms.
>>>> You could also let these other algorithms compute their routes
>>>> externally and then alter the routes in the simulation (via TraCI or xml
>>>> inputs).
>>>> regards,
>>>> Jakob
>>>>
>>>> Am Mo., 24. Feb. 2020 um 02:58 Uhr schrieb Raheleh Zarei <
>>>> [email protected]>:
>>>>
>>>>> Thanks for your explanation.
>>>>> Yes, you are right. So, what you suggest to do if I want to compare
>>>>> other routing algorithms with SUMO routing algorithms from traffic
>>>>> efficiency point of view?
>>>>> Do I need to develop SUMO source code and add other routing
>>>>> algorithms, or there is another way I can call these algorithms inside 
>>>>> SUMO.
>>>>>
>>>>> Thanks for your time and help.
>>>>> Raheleh
>>>>>
>>>>> On Tue, Feb 18, 2020 at 3:57 PM Jakob Erdmann <[email protected]>
>>>>> wrote:
>>>>>
>>>>>> Libsumo is useful if you need at lot of interaction between your code
>>>>>> and sumo because it reduce communication overhead compared to TraCI.
>>>>>> If you are comparing the algorithms in terms of overall traffic
>>>>>> efficiency then most of the time is spent on simulation rather than 
>>>>>> routing
>>>>>> so you won't need Libsumo.
>>>>>> If you are repeatedly calling the sumo routers to compare the found
>>>>>> routes with your own routes then it makes more sense.
>>>>>> For details, see https://sumo.dlr.de/docs/Libsumo.html
>>>>>> regards,
>>>>>> Jakob
>>>>>>
>>>>>> Am Mo., 17. Feb. 2020 um 10:19 Uhr schrieb Raheleh Zarei <
>>>>>> [email protected]>:
>>>>>>
>>>>>>> Hello,
>>>>>>>
>>>>>>> I would like to test two routing algorithms using SUMO as a
>>>>>>> framework for comparing these algorithms. Inside the program of these
>>>>>>> algorithms (written in C++ in Linux), I need to call SUMO, start it, get
>>>>>>> the required variables and reports and then close it and see the 
>>>>>>> results.
>>>>>>> What do you suggest for this? Do you think using Libsumo is a good idea?
>>>>>>> and if yes what's the procedure of using it.
>>>>>>>
>>>>>>> Best,
>>>>>>> Raheleh
>>>>>>> --
>>>>>>> *RAZ*
>>>>>>> _______________________________________________
>>>>>>> sumo-user mailing list
>>>>>>> [email protected]
>>>>>>> To change your delivery options, retrieve your password, or
>>>>>>> unsubscribe from this list, visit
>>>>>>> https://www.eclipse.org/mailman/listinfo/sumo-user
>>>>>>>
>>>>>> _______________________________________________
>>>>>> sumo-user mailing list
>>>>>> [email protected]
>>>>>> To change your delivery options, retrieve your password, or
>>>>>> unsubscribe from this list, visit
>>>>>> https://www.eclipse.org/mailman/listinfo/sumo-user
>>>>>>
>>>>>
>>>>>
>>>>> --
>>>>> *RAZ*
>>>>> _______________________________________________
>>>>> sumo-user mailing list
>>>>> [email protected]
>>>>> To change your delivery options, retrieve your password, or
>>>>> unsubscribe from this list, visit
>>>>> https://www.eclipse.org/mailman/listinfo/sumo-user
>>>>>
>>>> _______________________________________________
>>>> sumo-user mailing list
>>>> [email protected]
>>>> To change your delivery options, retrieve your password, or unsubscribe
>>>> from this list, visit
>>>> https://www.eclipse.org/mailman/listinfo/sumo-user
>>>>
>>>
>>>
>>> --
>>> *RAZ*
>>> _______________________________________________
>>> sumo-user mailing list
>>> [email protected]
>>> To change your delivery options, retrieve your password, or unsubscribe
>>> from this list, visit
>>> https://www.eclipse.org/mailman/listinfo/sumo-user
>>>
>> _______________________________________________
>> sumo-user mailing list
>> [email protected]
>> To change your delivery options, retrieve your password, or unsubscribe
>> from this list, visit
>> https://www.eclipse.org/mailman/listinfo/sumo-user
>>
>
>
> --
> *RAZ*
> _______________________________________________
> sumo-user mailing list
> [email protected]
> To change your delivery options, retrieve your password, or unsubscribe
> from this list, visit
> https://www.eclipse.org/mailman/listinfo/sumo-user
>
_______________________________________________
sumo-user mailing list
[email protected]
To change your delivery options, retrieve your password, or unsubscribe from 
this list, visit
https://www.eclipse.org/mailman/listinfo/sumo-user

Reply via email to