Machine learning prediction of late deliveries of orders placed with an online system

A computer program product comprising a non-transitory computer readable storage medium having instructions encoded thereon that, when executed by a processor, cause the processor to perform steps comprising:

• receiving, from a plurality of devices associated with a plurality of users of an online system and via a network, order data related to a plurality of orders placed by the plurality of users at the online system, the order data including real time information about a number of visits of a page of an application of the online system running on the plurality of devices within a time period, a number of currently unassigned orders of the plurality of orders, and a number of completed orders of the plurality of orders within the time period;
• receiving, from at least one of the plurality of devices associated with the plurality of users or a plurality of devices associated with a plurality of pickers and via the network, system state signals including real time information about a supply state of the online system and a demand state of the online system;
• accessing a delivery prediction machine-learning model of the online system, wherein the delivery prediction machine-learning model is trained to predict, in real time, a set of rates of late deliveries for a set of service options for delivery of the plurality of orders;
• applying the delivery prediction machine-learning model to the system state signals and the order data including the number of visits of the page of the application within the time period, the number of currently unassigned orders and the number of completed orders within the time period to generate, in real time, a set of late delivery scores, each late delivery score of the set of late delivery scores indicative of a respective rate of the set of rates of late deliveries for a respective service option of the set of service options for delivery of the plurality of orders;
• comparing each late delivery score of the set of late delivery scores to a respective threshold score of a set of threshold scores;
• responsive to each late delivery score being greater than the respective threshold score, triggering an action in relation to the respective service option for delivery of the plurality of orders;
• generating, in response to the triggered action and using each late delivery score, one or more action signals;
• sending, via the network, the one or more action signals to each device of the plurality of devices associated with each user of the plurality of users, wherein the sending causes a respective user interface of each device to display information about one or more service options of the set of service options for delivery of a respective order of the plurality of orders;
• receiving, via the network and from a device of the plurality of devices associated with a user of the plurality of users, information about the user placing an order of the plurality of orders including a set of items and the user selecting a service option of the one or more service options for delivery of the order;
• responsive to the user placing the order and selecting the service option, assigning a servicing of the order to a picker who is a semi-autonomous robot or a fully-autonomous robot;
• upon assigning the servicing of the order, instructing the picker operating as the semi-autonomous robot or the fully-autonomous robot to collect the set of items in a retailer location and deliver the set of items to the user by using an autonomous vehicle;
• physically collecting, by the picker operating as the semi-autonomous robot or the fully-autonomous robot, the set of items in the retailer location;
• and
• delivering, by the picker operating as the semi-autonomous robot or the fully-autonomous robot, the set of items to the user by using the autonomous vehicle.

A computer system comprising:

• a processor;
• and
• a non-transitory computer-readable storage medium having instructions that, when executed by the processor, cause the computer system to perform steps comprising:
• receiving, from a plurality of devices associated with a plurality of users of an online system and via a network, order data related to a plurality of orders placed by the plurality of users at the online system, the order data including real time information about a number of visits of a page of an application of the online system running on the plurality of devices within a time period, a number of currently unassigned orders of the plurality of orders, and a number of completed orders of the plurality of orders within the time period;
• receiving, from at least one of the plurality of devices associated with the plurality of users or a plurality of devices associated with a plurality of pickers and via the network, system state signals including real time information about a supply state of the online system and a demand state of the online system;
• accessing a delivery prediction machine-learning model of the online system, wherein the delivery prediction machine-learning model is trained to predict, in real time, a set of rates of late deliveries for a set of service options for delivery of the plurality of orders;
• applying the delivery prediction machine-learning model to the system state signals and the order data including the number of visits of the page of the application within the time period, the number of currently unassigned orders and the number of completed orders within the time period to generate, in real time, a set of late delivery scores, each late delivery score of the set of late delivery scores indicative of a respective rate of the set of rates of late deliveries for a respective service option of the set of service options for delivery of the plurality of orders;
• comparing each late delivery score of the set of late delivery scores to a respective threshold score of a set of threshold scores;
• responsive to each late delivery score being greater than the respective threshold score, triggering an action in relation to the respective service option for delivery of the plurality of orders;
• generating, in response to the triggered action and using each late delivery score, one or more action signals;
• sending, via the network, the one or more action signals to each device of the plurality of devices associated with each user of the plurality of users, wherein the sending causes a respective user interface of each device to display information about one or more service options of the set of service options for delivery of a respective order of the plurality of orders;
• receiving, via the network and from a device of the plurality of devices associated with a user of the plurality of users, information about the user placing an order of the plurality of orders including a set of items and the user selecting a service option of the one or more service options for delivery of the order;
• responsive to the user placing the order and selecting the service option, assigning a servicing of the order to a picker who is a semi-autonomous robot or a fully-autonomous robot;
• upon assigning the servicing of the order, instructing the picker operating as the semi-autonomous robot or the fully-autonomous robot to collect the set of items in a retailer location and deliver the set of items to the user by using an autonomous vehicle;
• physically collecting, by the picker operating as the semi-autonomous robot or the fully-autonomous robot, the set of items in the retailer location;
• and
• delivering, by the picker operating as the semi-autonomous robot or the fully-autonomous robot, the set of items to the user by using the autonomous vehicle.