The online pickup and delivery problem is motivated by the takeaway order delivery on crowdsourcing delivery platform, which is a newly emerged online to offline business model based on sharing economy. Considering the features of crowdsourcing delivery, an online pickup and delivery problem with constrained capacity is proposed, whose objective is to route a delivery man with constrained capacity to serve requests released over time so as to minimize the total latency. We consider online point-to-point requests with single pickup location where each request has to be picked up at the single pickup location and delivered to its destination, and each request become available at its release time, which is not known in advance. The lower bound of this problem for various capacities is proved. Two online algorithms WR and WI are presented, the competitive ratios on a half line and on general metric space are proved respectively. Further, a computational study is conducted to compare the performance of these two online algorithms on random instances of general metric space. The result shows algorithm WR performs better than WI in random cases but not in the worst case.

在线收货和送货问题是由众包交付平台上的外卖订单交付引起的，众包交付平台是一种基于共享经济的新兴在线到离线商业模式。考虑到众包交付的特征，提出了一种容量受限的在线取货和配送问题，其目的是路由容量受限的交付人服务随时间推移释放的请求，以最大程度地减少总延迟。我们考虑具有单个取件位置的在线点对点请求，其中每个请求都必须在单个取件位置被取走并传递到其目的地，并且每个请求在其发布时间就可用，这是事先未知的。证明了该问题对于各种容量的下限。提出了两种在线算法WR和WI，分别证明了半线和一般度量空间上的竞争比率。此外，进行了一项计算研究，以比较这两种在线算法在一般度量空间的随机实例上的性能。结果表明，算法WR在随机情况下的性能优于WI，但在最坏情况下则不如WI。