Generating recommendations for pickers servicing orders placed with an online concierge system based on actual and forecasted orders

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:

• receive location information associated with a plurality of pickers available to service orders placed with an online concierge system;
• receive a set of actual orders to be serviced by one or more pickers, wherein each actual order of the set of actual orders is associated with a geographical zone;
• access a machine learning model that is trained to predict a likelihood that an actual order associated with the geographical zone will be available for servicing within a timeframe;
• for each forecasted order of a set of forecasted orders associated with the geographical zone, apply the machine learning model to predict the likelihood that an actual order will be available for servicing within the timeframe;
• match each picker of the plurality of pickers to an order for servicing by minimizing a value of a matching and positioning function, wherein the matching and positioning function is based at least in part on:
• a first difference between a location associated with each picker matched to an actual order for servicing and a first retailer location associated with a corresponding actual order, a second difference between the location associated with each picker matched to a forecasted order for servicing and a second retailer location associated with a corresponding forecasted order, and the predicted likelihood for each forecasted order;
• generate one or more recommendations to present to each picker of the plurality of pickers based at least in part on the matching, the one or more recommendations for one or more of:
• accepting one or more actual orders for servicing, moving to one or more retailer locations associated with one or more forecasted orders within the timeframe, and checking back with the online concierge system at a later time for a set of actual and forecasted orders for servicing;
• and
• send the one or more recommendations for display to a picker client device associated with each picker of the plurality of pickers.

A computer system comprising:

• a processor;
• and
• a non-transitory computer readable storage medium storing instructions that, when executed by the processor, perform actions comprising:
• receiving location information associated with a plurality of pickers available to service orders placed with an online concierge system;
• receiving a set of actual orders to be serviced by one or more pickers, wherein each actual order of the set of actual orders is associated with a geographical zone;
• accessing a machine learning model that is trained to predict a likelihood that an actual order associated with the geographical zone will be available for servicing within a timeframe;
• for each forecasted order of a set of forecasted orders associated with the geographical zone, applying the machine learning model to predict the likelihood that an actual order will be available for servicing within the timeframe;
• matching each picker of the plurality of pickers to an order for servicing by minimizing a value of a matching and positioning function, wherein the matching and positioning function is based at least in part on:
• a first difference between a location associated with each picker matched to an actual order for servicing and a first retailer location associated with a corresponding actual order, a second difference between the location associated with each picker matched to a forecasted order for servicing and a second retailer location associated with a corresponding forecasted order, and the predicted likelihood for each forecasted order;
• generating one or more recommendations to present to each picker of the plurality of pickers based at least in part on the matching, the one or more recommendations for one or more of:
• accepting one or more actual orders for servicing, moving to one or more retailer locations associated with one or more forecasted orders within the timeframe, and checking back with the online concierge system at a later time for a set of actual and forecasted orders for servicing;
• and
• sending the one or more recommendations for display to a picker client device associated with each picker of the plurality of pickers.