Naturally, the whole thing works automatically and requires no human [usually cack-handed] interference, barring force majeure.
After a certain density of deliveries in sectors is reached, the subj ceases to be a trivial task, and reliance on the route’s obviousness and the driver’s common sense is less and less justified by practical results.
Whereas hidden optimization resources are significant: 22 or 24 daily deliveries per vehicle make a big difference.
Especially if you have dozens or hundreds of vehicles.
In general terms, the travelling salesman problem is a very complicated mathematical problem. But in our narrowly formulated conditions it is quite easy to solve using computerized algorithms.
There are specialized logistical software products, of which Antor is one of the best-known.
Its work scheme is simple: your list of deliveries is uploaded into Antor in a certain format and then downloaded back, sorted according to the optimal order of execution:
Antor is rather a powerful piece of software with sprawling functionality.
The underside is that it is rather expensive and hard to set up and use — too much so, given the limited tasks of the automation of urban distribution.
Also on the negative side, the lack of automatic data exchange (Antor had no API implemented by that time) necessitated manual upload and download by the operator. With all the entailments.
In short, after having worked with Antor for some time, at least three dia$par operator companies renounced it.
For those customers we developed our own optimal route calculation service, naturally integrated into dia$par.
The routes are calculated using the Google and Yandex API s that provide convenient and efficient geocoding and geolocation tools.
And, what is damn specific, perfectly free ones.
The Google and Yandex APIs are used to extract the coordinates of the delivery points and the actual motorway distances between the pairs of addresses.
The optimal routes are calculated in dia$par with the time intervals, the goods" weight and dimensions, etc. taken into account.