Spatial Two-Sided Online Bottleneck Matching with Deadlines

Recently, there are several studies focusing on the bottleneck optimization objective in Spatial Crowdsourcing (SC). However, these studies usually do not consider the deadline constraint. Different from these studies, we take deadlines into consideration and identify the Fully Online Bottleneck Matching with Deadlines (FOBMD) problem in SC. Because of the deadlines, consideration must be given to both the bottleneck cost and the cardinality, which makes the FOBMD problem more challenging, and no online algorithm without actively refusing tasks can achieve a constant competitive ratio of the bottleneck cost for the FOBMD problem. To settle the FOBMD problem, we consider three baseline algorithms and propose an online algorithm, namely Local Isolated Point Greedy (LIPG). Finally, we validate the effectiveness and efficiency of our proposed algorithm via extensive experiments on both synthetic and real world datasets.