Decision models for personal shopper platform operations optimization

Abstract

The ‘lazy economy’ gives rise to an emerging business mode, called personal shopper platforms (PSPs). A customer who needs some goods urgently can release an order on a PSP, which is then assigned by the platform to a personal shopper, who will buy the goods at a nearby retail store and deliver them to the customer within a short time interval. Since the development of PSP is relatively new, the decision mechanisms and policies are at an early stage. The operations of the PSPs can be optimized through operations research methodologies. This study proposes a series of mixed integer programming (MIP) models and improved dynamic programming-based algorithms to support operational decisions on order assignment and shopper routing, as well as strategic decisions on the PSP mode adoption and territory planning. Some intuitive but practical criteria are also designed to accelerate the proposed algorithms so that they can be applied to large-scale realistic instances. The proposed algorithm can solve the basic case with 1000 orders and 1050 shoppers (about 107 variables and 107 constraints in the MIP models) in half a minute. A realistic case in the Changning district of Shanghai is also used to validate the effectiveness of the proposed models and the efficiency of the algorithms. An extended model for considering the uncertain arrival of future orders is also presented. This study provides a comprehensive model-driven decision methodology for this emerging service industry mode.