Partitioning Orders in Online Shopping Services
Abstract
The rapid growth of the sharing economy has led to the widespread use of newer and richer models of online shopping and delivery services. The race to deliver fast has transformed such services into complex networks of shoppers,
stores, and consumers. Needless to say, the efficiency of the store order management is critical to the business.
Motivated by this setting, we consider the following problem: given a set of online shopping orders each consisting of a few items, how to best partition the orders among a given number of pickers? Owing to logistical constraints the orders are typically unsplittable in the partition. This partitioning, taking the physical location of the items in the store , has to optimize the utilization and amount of work done by the shoppers in the store. Formulating this as a combinatorial optimization problem, we propose a family of simple and efficient algorithms that admit natural constraints arising in this setting. In addition to showing provable guarantees for the algorithms, we also demonstrate their efficiency in practice on real-world data from Google Express [1], outperforming natural baselines.
stores, and consumers. Needless to say, the efficiency of the store order management is critical to the business.
Motivated by this setting, we consider the following problem: given a set of online shopping orders each consisting of a few items, how to best partition the orders among a given number of pickers? Owing to logistical constraints the orders are typically unsplittable in the partition. This partitioning, taking the physical location of the items in the store , has to optimize the utilization and amount of work done by the shoppers in the store. Formulating this as a combinatorial optimization problem, we propose a family of simple and efficient algorithms that admit natural constraints arising in this setting. In addition to showing provable guarantees for the algorithms, we also demonstrate their efficiency in practice on real-world data from Google Express [1], outperforming natural baselines.
Research Areas
