Model-Based Rolling Matching Strategy for Crowdsourced Drivers and Delivery Tasks Considering Uncertain Transportation Duration

Abstract

The matching of crowdsourced drivers and delivery tasks is an important decision problem for the crowdsourced delivery platform. Although the existence of uncertainty of transportation duration in logistics delivery has been verified, uncertain transportation duration has not been considered in previous studies on the matching of crowdsourced drivers and delivery tasks. This would lead to the limitation that the results of the existing methods cannot meet the time requirements of senders. In this case, the profit and customer satisfaction of the crowdsourced delivery platform would decrease. In this paper, a model-based rolling matching strategy to match crowdsourced drivers and delivery tasks considering uncertain transportation duration is proposed. In addition, it is assumed that the crowdsourced delivery platform also has some dedicated drivers to implement the delivery tasks that cannot be implemented by crowdsourced drivers. First, a simpler problem is described, which is to match crowdsourced drivers and delivery tasks considering uncertain transportation duration in a static data environment. Then, a model is proposed to solve the above problem. Based on the proposed model, this paper further proposes a rolling procedure to solve the problem in a data refreshing environment. Moreover, a heuristic algorithm is presented for combining multiple delivery tasks to solve the one-to-many matching. Finally, a case study and comparison are given to illustrate the validity and the contribution of the proposed matching strategy. The results show that the proposed matching strategy has a distinct advantage of cost savings.