Automatic generation of dynamic time-slot capacity

Abstract

A system including one or more processors and one or more non-transitory computer-readable media storing computing instructions that, when executed on the one or more processors, cause the one or more processors to perform certain acts. The acts can include generating, based on a trained machine learning model, one or more time-slot capacities for one or more pickup time slots at a physical store for a time period that has not yet occurred. The acts also can include, after the time period has occurred, determining when actual demand exceeded the one or more time-slot capacities to tune the trained machine learning model. Other embodiments are described.

What is claimed is: 
     
         1 . A system comprising:
 one or more processors; and   one or more non-transitory computer-readable media storing computing instructions that, when executed on the one or more processors, cause the one or more processors to perform:
 generating, based on a trained machine learning model, one or more time-slot capacities for one or more pickup time slots at a physical store for a time period that has not yet occurred; and 
 after the time period has occurred, determining when actual demand exceeded the one or more time-slot capacities to tune the trained machine learning model. 
   
     
     
         2 . The system of  claim 1 , wherein the trained machine learning model comprises a recurrent neural network model comprising a long short-term memory cell. 
     
     
         3 . The system of  claim 1 , wherein the computing instructions, when executed on the one or more processors, cause the one or more processors to further perform, before generating the one or more time-slot capacities:
 training a machine learning model to create the trained machine leaning model, based on historical demand data for the pickup time slots over a time period that already occurred, wherein:
 the historical demand data is represented as a set of input vectors as input to the machine learning model; and 
 each input vector of the set of input vectors represents a demand at the physical store for a single day in the time period that already occurred. 
   
     
     
         4 . The system of  claim 3 , wherein the time period that already occurred is approximately four weeks. 
     
     
         5 . The system of  claim 4 , wherein each input vector of the set of input vectors comprises a set of elements each representing a number of pickups that were scheduled at the physical store for a respective one of pickup time slots that occurred during the single day. 
     
     
         6 . The system of  claim 1 , wherein the computing instructions, when executed on the one or more processors, cause the one or more processors to further perform, before determining when the actual demand exceeded the one or more time-slot capacities:
 presenting the one or more pickup time slots to a user who requests a pickup during the time period that has not yet occurred, based on the one or more time-slot capacities.   
     
     
         7 . The system of  claim 1 , wherein the one or more time-slot capacities represents a maximum number of pickups that can be scheduled with each of one or more time slots at the physical store for the time period that has not yet occurred. 
     
     
         8 . The system of  claim 7 , wherein the maximum number of pickups comprises a maximum combined number of customer pickups and delivery pickups at the physical store. 
     
     
         9 . The system of  claim 1 , wherein generating the one or more time-slot capacities further comprises:
 generating the one or more time-slot capacities based on predetermined capacity range constraints.   
     
     
         10 . The system of  claim 1 , wherein the one or more time-slot capacities are used for scheduling a number of workers at the physical store for each day of the time period that has not yet occurred. 
     
     
         11 . A method being implemented via execution of computing instructions configured to run at one or more processors and stored at one or more non-transitory computer-readable media, the method comprising:
 generating, based on a trained machine learning model, one or more time-slot capacities for one or more pickup time slots at a physical store for a time period that has not yet occurred; and   after the time period has occurred, determining when actual demand exceeded the one or more time-slot capacities to tune the trained machine learning model.   
     
     
         12 . The method of  claim 11 , wherein the trained machine learning model comprises a recurrent neural network model comprising a long short-term memory cell. 
     
     
         13 . The method of  claim 11  further comprising, before generating the one or more time-slot capacities:
 training a machine learning model to create the trained machine leaning model, based on historical demand data for the pickup time slots over a time period that already occurred, wherein:
 the historical demand data is represented as a set of input vectors as input to the machine learning model; and 
 each input vector of the set of input vectors represents a demand at the physical store for a single day in the time period that already occurred. 
 
 
     
     
         14 . The method of  claim 13 , wherein the time period that already occurred is approximately four weeks. 
     
     
         15 . The method of  claim 14 , wherein each input vector of the set of input vectors comprises a set of elements each representing a number of pickups that were scheduled at the physical store for a respective one of pickup time slots that occurred during the single day. 
     
     
         16 . The method of  claim 11  further comprising, before determining when the actual demand exceeded the one or more time-slot capacities:
 presenting the one or more pickup time slots to a user who requests a pickup during the time period that has not yet occurred, based on the one or more time-slot capacities. 
 
     
     
         17 . The method of  claim 11 , wherein the one or more time-slot capacities represents a maximum number of pickups that can be scheduled with each of one or more time slots at the physical store for the time period that has not yet occurred. 
     
     
         18 . The method of  claim 17 , wherein the maximum number of pickups comprises a maximum combined number of customer pickups and delivery pickups at the physical store. 
     
     
         19 . The method of  claim 11 , wherein generating the one or more time-slot capacities further comprises:
 generating the one or more time-slot capacities based on predetermined capacity range constraints.   
     
     
         20 . The method of  claim 11 , wherein the one or more time-slot capacities are used for scheduling a number of workers at the physical store for each day of the time period that has not yet occurred.