System and method for processing shipment requests using a multi-service shipping platform

Abstract

Systems and methods for processing shipment request by using a multi-carrier shipping services platform.

We claim: 
     
         1 . A system, comprising:
 an interface configured to receive a first location and a second location associated with a new package;   a time-in-transit module configured to, for each of a set of services:
 for each of a set of lanes:
 generate bi-directional training data comprising aggregated historical data associated with the service across both directions of the lane, wherein the lane is determined based on the first location and the second location; 
 generate a model based on the bi-directional training data; 
 
 select a model associated with a lane, wherein the bi-directional training data associated with the selected model comprises statistically significant bi-directional training data; and 
 for each percentile in a set of percentiles, determine a time-in-transit for the percentile of packages to be delivered, using the selected model; 
   
       wherein the interface returns the time-in-transits for one or more services from the set of services. 
     
     
         2 . The system of  claim 1 , wherein each lane is determined based on a partial address for at least one of the first location or the second location. 
     
     
         3 . The system of  claim 2 , wherein the partial address for each lane is determined such that the aggregated historical data across both directions of the lane comprises at least a threshold number of historical packages. 
     
     
         4 . The system of  claim 2 , wherein each lane is determined based on a partial address for the first location at a first level of granularity and a partial address for the second location at a second level of granularity. 
     
     
         5 . The system of  claim 2 , wherein each lane is associated with historical packages across both directions of the lane, wherein the lane corresponding to the selected model for each service is associated with a minimum number of historical packages above a threshold. 
     
     
         5 . The system of  claim 2 , wherein the lane corresponding to the selected model for each service is defined at a minimum level of granularity that satisfies a threshold requirement for a number of historical packages associated with both directions of the lane. 
     
     
         6 . The system of  claim 1 , wherein the model is selected for each of the set of services based on the number of historical packages associated with the corresponding lane. 
     
     
         7 . The system of  claim 1 , wherein, for each of the set of services, the time-in-transit is concurrently determined for each of the set of percentiles. 
     
     
         8 . The system of  claim 1 , wherein each model is a predictive neural network model trained on the bi-directional training data. 
     
     
         9 . The system of  claim 1 , wherein uni-directional data, comprising aggregated historical data across a single direction of the lane corresponding to the selected model, is not statistically significant for each of the set of services. 
     
     
         10 . The system of  claim 1 , wherein the interface is further configured to receive a percentile requirement and a time-in-transit requirement for the new package, wherein the time-in-transit module is further configured to select the one or more services that have respective time-in-transits for the percentile requirement that are no more than the time-in-transit requirement. 
     
     
         11 . The system of  claim 1 , wherein the time-in-transit module is further configured to select the set of percentiles such that each percentile in the set is greater than or equal to a percentile requirement associated with the new package. 
     
     
         12 . A method, comprising:
 for each service in a set of services, automatically:
 for each lane in a set of lanes, aggregating historical transit data for packages associated with the service across both directions of the lane to generate bi-directional data, wherein the lane is determined based on a first location and a second location; 
 selecting a lane from the set of lanes, wherein the bi-directional training data associated with the selected lane comprises statistically significant bi-directional training data; 
 for each time-in-transit in a set of time-in-transits, determining a percentile of packages delivered within the time-in-transit, based on the statistically significant bi-directional data for the selected lane; 
   selecting a subset of the set of services for a new package based on the percentile for each time-in-transit; and   displaying each time-in-transit and the corresponding percentiles for the subset of services.   
     
     
         13 . The method of  claim 12 , wherein determining the percentile of packages delivered within the time-in-transit comprises:
 training a model based on the statistically significant bi-directional data; and   determining the percentile using the model.   
     
     
         14 . The method of  claim 13 , wherein the model is a neural network model trained on historical percentiles of packages delivered within the time-in-transit, wherein the percentile is a predicted percentile for the new package. 
     
     
         15 . The method of  claim 12 , wherein determining the percentile of packages delivered within the time-in-transit is further based on package information associated with the new package, wherein the package information comprises at least one of a package weight or a package dimension. 
     
     
         16 . The method of  claim 12 , further comprising selecting the set of time-in-transits such that each time-in-transit in the set is less than or equal to a time-in-transit requirement associated with the new package. 
     
     
         17 . The method of  claim 12 , wherein selecting the subset of the set of services comprises: selecting services associated with a determined percentile of packages for a time-in-transit requirement, wherein the determined percentile of packages is greater than or equal to a percentile requirement, wherein the time-in-transit requirement and the percentile requirement are associated with the new package. 
     
     
         18 . The method of  claim 12 , wherein uni-directional data comprising aggregated historical data for packages across a single direction of the selected lane is not statistically significant for each of the set of services. 
     
     
         19 . The method of  claim 12 , wherein each lane is determined based on a partial address for at least one of the first location or the second location, wherein the selected lane corresponds to aggregated historical data across both directions of the lane comprising at least a threshold number of historical packages. 
     
     
         20 . The method of  claim 19 , wherein each lane is determined based on a partial address for the first location at a first level of granularity and a partial address for the second location at a second level of granularity.