Courier routing for a new last-mile logistics service
研究即时配送平台“帮我买”服务的运营优化，提出两阶段随机规划模型和原始分解算法，帮助平台决定是否接单、分配订单和选择商店，以最大化利润。
As a new business mode for last-mile logistics services, some instant delivery platforms provide “help me buy” services mainly to satisfy urgent customer demands. Couriers travel to nearby stores, buy commodities requested by the customer, and quickly deliver them to the customer’s location. We investigate how to operate this type of platform to maximize profits. A two-stage stochastic programming model is proposed to determine whether to accept a customer’s order, how to assign accepted orders to couriers, and how to select stores where the commodities are purchased. The proposed model can be applied using a rolling horizon approach and account for the uncertain arrival of future orders in each epoch. Moreover, because the second-stage subproblem involves an integer programming model, a new primal decomposition algorithm is proposed to solve the two-stage model. Extensions are also explored so that our proposed methodology can be applied to more general and realistic platform operations. Numerical experiments based on real data are conducted to test the effectiveness of the proposed algorithm and to derive useful managerial insights for operators of help me buy services. This study indicates the necessity of applying the new decomposition to this new delivery model, considering future orders, and allowing platform collaboration, and determines the benefit of the stochastic solution. In addition, this study reveals the influences of the courier-to-order ratio; response time; distribution and radius of order, courier, and store circles; demand density; and decision frequency.