I’m working on a project with https://207inc.jp. In this case, the most efficient route is found by mixing timed and untimed packages for a clustering-optimized delivery destination. This allows you to choose the route that will be the most efficient in total for all time periods.
In this case, we design to incorporate the delivery of untimed parcels into the three timed parcels and the delivery of untimed parcels into the timed parcels. In this case, we introduce a parameter that assumes an approximate distance; the distribution of the three time zones corresponds to yellow, red and blue, and the unspecified time zone corresponds to green.
After looking at the overall balance and re-categorizing the green into each time zone with an overall optimization, the result is
Time Zone 2 was originally a lot of deliveries, so the allocation was reduced. Time Zone 1 was less, so there were more new assignments. This time, we used D-Wave’s Leap2 to do the calculations and were able to efficiently distribute untimed packages.