US 12,462,216 B1
Enhanced graph-based modeling for geographic region optimization
Julie Marie Paule Poullet, Seattle, WA (US); Andre Cornelis Joseph Snoeck, Austin, TX (US); Daniel Esteban Merchan Duenas, Austin, TX (US); Marc Joseph Anderson, Driftwood, TX (US); Julian Enrique Pachon, Austin, TX (US); Shalini Vangala, Mount Juliet, TN (US); Mayur Rajendra Giri, Nashville, TN (US); Raj Kumar Jayapalan, Lago Vista, TX (US); and Devender Kumar Gunti Venkata, Redmond, WA (US)
Assigned to Amazon Technologies, Inc., Seattle, WA (US)
Filed by Amazon Technologies, Inc., Seattle, WA (US)
Filed on Sep. 22, 2023, as Appl. No. 18/473,001.
Int. Cl. G06Q 10/0835 (2023.01); G01C 21/00 (2006.01); G01C 21/34 (2006.01)
CPC G06Q 10/08355 (2013.01) [G01C 21/3446 (2013.01); G01C 21/3867 (2020.08)] 20 Claims
OG exemplary drawing
 
1. A method for graph-based modeling to optimize geographic regions using a resource-constrained objective-maximal path, the method comprising:
identifying, by at least one processor of a device, a change in jurisdiction for a delivery service with multiple delivery service providers from a first jurisdiction comprising a first set of geographic areas to a second jurisdiction comprising a second set of geographic areas;
inputting, by the at least one processor, geographic districts, of the second set of geographic areas, comprising first geographic districts for a first delivery service provider (DSP) of the delivery service and second geographic districts for a second DSP of the delivery service;
mapping, by the at least one processor, a first geographic district of the first geographic districts to the first DSP based on delivery routes per day of the first geographic district and delivery routes per day of the first DSP;
mapping, by the at least one processor, a second geographic district and a third geographic district of the second geographic districts to the second DSP based on delivery routes per day of the second geographic district and delivery routes per day of the second DSP;
generating, by the at least one processor, based on the mapping for the first geographic district, a first deviation between the delivery routes per day of the first geographic district and the delivery routes per day of the first DSP, a first geographically contiguous geographic area for the first DSP, the first geographically contiguous geographic area overlapping at most two geographic districts;
generating, by the at least one processor, based on the mapping for the second geographic district, a second deviation between the delivery routes per day of the second geographic district and the delivery routes per day of the second DSP, a second geographically contiguous geographic area for the second DSP, the second geographically contiguous geographic area overlapping at most two geographic district;
encoding, by the at least one processor, the first geographically contiguous geographic area as a first path between a first origin location and a first destination location, the first path comprising a first vertex of the first geographic district and the first DSP, and comprising a first arc between two vertices of the first path;
encoding, by the at least one processor, the second geographically contiguous geographic area as a second path between a second origin location and a second destination location, the second path comprising a second vertex of the second geographic district and the second DSP, and comprising a second arc between two vertices of the second path;
applying, by the at least one processor, a set of constraints to the first geographically contiguous geographic area and the second geographically contiguous geographic area to cover the second jurisdiction with geographically contiguous geographic areas, and to map respective DSPs to a single respective geographically contiguous geographic area;
generating, by the at least one processor, a first objective-maximal path between the first origin location and the first destination location based on an upper bound of the delivery routes per day of the first geographic district and a lower bound of the delivery routes per day of the first DSP; and
generating, by the at least one processor, a second objective-maximal path between the second origin location and the second destination location based on an upper bound of the delivery routes per day of the second geographic district and a lower bound of the delivery routes per day of the second DSP,
wherein the first objective-maximal path maximizes a first familiarity of the DSP with the first objective-maximal path based on historically delivered volumes by the DSP to the first geographic district in a previous time window, and
wherein the second objective-maximal path maximizes a second familiarity of the DSP with the second objective-maximal path based on historically delivered volumes by the DSP to the second geographic district in the previous time window.