Tour Optimization: using Drones

This forum deals with any kind of trip optimization whether it is automatic planning or manual dispatching, refering to transport orders or service planning.
Post Reply
User avatar
Bernd Welter
Site Admin
Posts: 2574
Joined: Mon Apr 14, 2014 10:28 am
Contact:

Tour Optimization: using Drones

Post by Bernd Welter »

Cheers,

these days I've been asked whether and how a tour planning call could consider a fleet that also uses drones.
Here are my thoughts / clarifications:
  • The main aspect of a drone in my eyes is that it simply performs it's distances based on airline with a static flight speed. So if you simply want to plan with DirectDistance you should keep in mind that the purpose of such an approach is a combination of an approximated detour where you would have to set the factor to 1.0 in combination with the estimated flight speed.
  • Therefore I'd recommend to use a routed distance matrix (ExistingDistanceMatrix) which is based on the built in fallback to geometric distance. For this you could apply a routing profile which "fails by design", e.g. by setting a ridiculous geographic restriction (meaningless bounding rectangle, wrong allowed_countries).
  • Such a distance matrix allows you to assemble a fleet based on regular vehicles (sprinters, cars...) and such drones via ExistingDistanceMatrixPerVehicle.
  • Keep in mind that drones probably have a maximum distance to serve due to battery capacities. Therefore the usage of a proper restriction such as Vehicle.MaximumDistance sounds meaningful.
  • As drones might come into use as some kind of a local satellite in the last mile context please be aware that the PTV xTour2 does not consider to handle the flow of an order via more than a single vehicle. You could try to modell this anyway with iterative strategies and business logic such as described in this article about the search for parking spots.
  • PS: for the usage of the simple routing API there's no fallback - just use airline geometries to display drone flights on a map.
Please let me know if you are aware of further aspects of drone based optimization tasks. It will be a pleasure to provide feedback or hints.

Bernd
Attachments
A sample 4x4 matrix output: the waypoints are in Germany but the allowedCountries are restriced to Austria. Therefore the routings performed a fallback to airline. WHich can be seen in the "estimatedByDirectDistance" column
A sample 4x4 matrix output: the waypoints are in Germany but the allowedCountries are restriced to Austria. Therefore the routings performed a fallback to airline. WHich can be seen in the "estimatedByDirectDistance" column
Bernd Welter
Technical Partner Manager Developer Components
PTV Logistics - Germany

Bernd at... The Forum,LinkedIn, Youtube, StackOverflow
I like the smell of PTV Developer in the morning... :twisted:
Joost
Posts: 307
Joined: Fri Apr 25, 2014 1:46 pm

Re: Tour Optimization: using Drones

Post by Joost »

Bernd Welter wrote:
  • Therefore I'd recommend to use a routed distance matrix (ExistingDistanceMatrix) which is based on the built in fallback to geometric distance. For this you could apply a routing profile which "fails by design", e.g. by setting a ridiculous geographic restriction (meaningless bounding rectangle, wrong allowed_countries).
Better is to pre-calculate the dima your self and use createDistanceMatrix with an ImportDistanceMatrixRequest payload to get it into the xServer. This gives you more control over distance and travel time calculation.

(Post: Import own distance matrix data)

For real world usage you also might need to take drone no-fly zones into account.
Joost Claessen
Senior Technical Consultant
PTV Benelux
Post Reply