Splitting Shopping and Delivery Tasks in an On-Demand Personal Shopper Service

Abstract

We introduce a personal shopper service that offers the same-day delivery of products available at brick and mortar stores. Each customer request is placed online and could involve items from multiple stores. The operator of this service has to dynamically accept requests, coordinate a fleet of shoppers, schedule shopping operations at stores and execute deliveries of accepted requests to customers on time. Our work focuses on studying the benefits of splitting requests into smaller tasks served by different shoppers in parallel and consolidating tasks with a common shopping location into one store visit. We present and solve an online optimization model and empirically show that the percentage of requests served increases when request splitting is allowed. This benefit is mostly obtained due to an increased shopper utilization, reduced shopping times, and cheaper routing options available.