Page 1 of 1

Manipulate DiMa distances/costs

Posted: Wed Mar 24, 2021 2:47 pm
by Wilhelm
Is there a possibility to manipulate values in a distance matrix to be used for xTour requests?
We would like to support handling jobs in areas without map material (e.g. city parks). With xServer2 the DiMa has to be calculated before using xTour. So, it would be nice if there is a possibility to manipulate "manually" the distances/costs for such jobs in the area without any streets. xTour should be able to handled this as it only uses the precalculated DiMa.

Re: Manipulate DiMa distances/costs

Posted: Fri Mar 26, 2021 8:02 am
by Bernd Welter
Hello Daniel,

we are aware of the need for such a feature. Most players use the combination of xDima and xTour.

Some partners build their own optimization based on specofic industries needs but use the xDima for the required distance matrices and some others - such as you - would like to push their own distances / Travel times (e.g. based on bus timetables) into xDima to provide a distance "metric" which is then used by xTour.

I forwarded your post to product management to let the colleagues know that another partner is interested in this feature!

Thanks for that,
Bernd

Re: Manipulate DiMa distances/costs

Posted: Fri Mar 26, 2021 12:22 pm
by Thorsten
Dear PTV-Team, ...like that! We do have this requirement since a while for countries with many islands (Greece) or very large countries RUS/USA/LATAM, where field-reps take ferry-boats or planes for certain relations and then switch back to cars etc. This is a good first step into the direction of inter-modal routing, although we expect PTV to deliver more sophisticated solutions soon. Question: Does this approach also work for xCluster (2.*)? THX&br Thorsten

Re: Manipulate DiMa distances/costs

Posted: Fri Mar 26, 2021 12:32 pm
by Bernd Welter
Hello Thorsten,

About the second question "Does this approach also work for xCluster (2.*)"

Yes, this would also work like this in xCluster 2. The concept of the DIMA handling in the xServer2 is based on three levels:
  • Use xData to create a High Performance Searchgraph (if possible and meaningful)
  • Use xDima to create the required dima structure(s) and consume the DIMA IDs (hope for a proper search graph)
  • Refer to the distance matrices via the IDs in the plpanning step (xTour, xCluster)
Of course you can also use the distance matrices for further purposes such as
  • Creation of reference tables between locations (e.g. from all your offices to all your customers)
  • Determination of the closest restaurants to your location (1:many)
  • Determination of the closest pizza service around you (many:1)
I recommend to look into these videos, too:

Bernd

PS: looks like I need a snack :D