Impact routing profile on tour quality...
Posted: Thu Dec 14, 2017 7:59 am
HI there,
as Lars and I have been challenged to optimize the structure of a sequencing topic I'd like to let you know what caused the poor quality
In the end it wasn't the optimization that failed but the distance matrices.
But obviously the cause for the issue wasn't the optimization. After zooming closer to the section we saw that this area is some kind of a pedestrian zone and so I decided to add routing afterwards (green line) and even the connection between the waypoints coordinates and the matchpoints (magenta).
That's it. Keep it in mind. Could be important for future inner city optimization issues.
Thanks to Lars for the hint with the linking...
Best regards,
Bernd
as Lars and I have been challenged to optimize the structure of a sequencing topic I'd like to let you know what caused the poor quality
In the end it wasn't the optimization that failed but the distance matrices.
But obviously the cause for the issue wasn't the optimization. After zooming closer to the section we saw that this area is some kind of a pedestrian zone and so I decided to add routing afterwards (green line) and even the connection between the waypoints coordinates and the matchpoints (magenta).
- Left: optimization and routing based on "car" profile
Right: both computed with a delivery profile.
As you can see about 20 of the magenta match points in the critical area are linked to the Kringstraße ans simply optimized based on the y-coordinate. So this explains the sequence.
After applying a delivery profile the linking distances are shortened and match the waypoints to the proper street.
Thanks to Lars for the hint with the linking...
Best regards,
Bernd