v.net.salesman and forward/backward movement cost

[lluis]

I'm doing some network analysis with GRASS GIS and I'm surprised because is not possible to consider forward/backward costs movements in v.net.salesman as it is possible with the v.net.path module. I would like to see this option in upcoming releases of GRASS because I believe that this enhancement could be so valuable for grass users

I'm working with streets of a city, so distinguish between one-way streets and bidirectional streets is something critical for me.

I'm working with grass 6.4.2svn under Ubuntu Congratulations for your work

Lluís

[mmetz]

Implemented in grass 7, r46229, please test!

Markus M

[mlennert]

Replying to mmetz:

Implemented in grass 7, r46229, please test!

Here are the results of an abritrary test on in the NC demo dataset:

​equal cost in both directions

​different cost in each direction for one way streets

I'll attach the script with the relevant commands I used.

At first view, the results seem to be coherent, but probably need more detailed testing.

Moritz

[mlennert]

commands used for testing

[Lluis]

Great! It seems to work properly. I've been doing some tests using open street map cartography, and it seems that I'm getting valid results. I will try to do more tests but it seems to work correctly.

Thanks a lot for your work and effort, Moritz!

Lluís

Replying to mlennert:

Replying to mmetz:

Implemented in grass 7, r46229, please test!

Here are the results of an abritrary test on in the NC demo dataset:

​equal cost in both directions

​different cost in each direction for one way streets

I'll attach the script with the relevant commands I used.

At first view, the results seem to be coherent, but probably need more detailed testing.

Moritz

[mlennert]

Replying to Lluis:

Thanks a lot for your work and effort, Moritz!

Markus did the work, I just did the testing, so I think the thanks should go to him. ;-)

I think this should be backported at least to grass6dev.

Moritz

[mmetz]

Replying to mlennert:

Replying to Lluis:

Thanks a lot for your work and effort, Moritz!

Markus did the work, I just did the testing, so I think the thanks should go to him. ;-)

Thanks to both of you for additional testing!

I think this should be backported at least to grass6dev.

OK.

BTW, there is also v.net.steiner using the same cost for both directions, no option to define different forward/backward moving costs.

Markus M

[mlennert]

Replying to mmetz:

Replying to mlennert:

Replying to Lluis:

Thanks a lot for your work and effort, Moritz!

Markus did the work, I just did the testing, so I think the thanks should go to him. ;-)

Thanks to both of you for additional testing!

I think this should be backported at least to grass6dev.

OK.

BTW, there is also v.net.steiner using the same cost for both directions, no option to define different forward/backward moving costs.

Another module where this might be an issue is v.net.components. Actually there is no cost parameter at all in this module, but strong or weak connection is defined by direction, and I don't understand how this direction is defined.

Moritz

[neteler]

What is the state here? G7 done and G6 wontfix?

[mlennert]

Replying to neteler:

What is the state here? G7 done and G6 wontfix?

I would definitely say wontfix for G6. For G7, v.net.steiner still does not have directions, so as a reminder maybe this bug should either be renamed, or a different bug opened for v.net.steiner.

[martinl]

Milestone renamed

[neteler]

Ticket retargeted after milestone closed

[martinl]

All enhancement tickets should be assigned to 7.6 milestone.

[martinl]

Ticket retargeted after milestone closed

[martinl]

Ticket retargeted after milestone closed