Joint workforce scheduling and pricing in crowdsourced delivery under uncertainty
研究了众包配送平台如何联合决定正式员工、预约配送员和临时配送员的排班、订单分配和补偿定价，以应对订单模式、配送员位置和路线时间的不确定性，实现成本节约。
The on-demand and same-day delivery market has seen rapid growth, and many delivery service platforms seek gig workers to supplement in-house couriers to fulfill surging orders in a timely fashion. The incorporation of gig delivery workers, referred to as ad hoc couriers, brings significant operational and tactical challenges as they are not directly controlled by the platform. Ad hoc couriers can decide whether to take an order based on the convenience of the delivery and the offered compensation price. To better manage the workforce, we model and solve a joint workforce scheduling and pricing problem (WSPP) with three types of labor, namely in-house couriers, scheduled couriers, and ad hoc couriers for last-mile delivery systems. We jointly decide the shifts of scheduled couriers, order assignment, and compensation pricing for ad hoc couriers, capturing the uncertainties associated with the delivery order pattern, the location of ad hoc couriers, and routing time. We propose an integrated approach that leverages stochastic programming, robust optimization, and machine learning to model the three types of uncertainties. To solve WSPP, we first derive structural properties of the model, based on which we devise an improved logic-based benders decomposition nested column-and-constraint generation algorithm with stepwise tightening relaxation. Then, we propose several acceleration techniques to speed up the solution process of large-scale instances. We validate the superior performance of the proposed model on real and synthetic data sets. Based on the out-of-sample evaluation, the proposed framework can deliver up to 10% cost savings over stochastic and deterministic methods. Moreover, it significantly reduces the volatility of the cost performance on test instances.