Dynamic Food Delivery Problem Based on Spatial Crowdsourcing

In recent years, crowdsourcing online food delivery (COFD) services have been increasingly popular, in which a number of crowdsourced riders are recruited to deliver food orders for those geographically dispersed customers. Due to the dynamics and uncertainty of food orders and riders, it is challenging to design an immediate allocation mechanism to recruit suitable riders and plan paths, with the goal of maximizing the total profit of all recruited riders. To address this challenge, we first formalize a crowdsourcing online food delivery problem with the goal of maximizing the expectation of long-term profit of all riders. Then, an efficient heuristic-based algorithm is proposed in which order-exchange and order-transfer policies are applied. Finally, we conduct extensive experiments on synthetic and real-world datasets to evaluate our proposed algorithm, whose results show that the proposed policies and algorithm are more effective compared to the baselines in terms of total profit, total distance, and average waiting time.