Real-time Multi-platform Route Planning in ridesharing

The increasing availability of mobile Internet and portable devices has led to the popularity of spatial crowdsourcing recently. The significant burgeoning of ridesharing services in spatial crowdsourcing has transformed urban mobility. However, due to the imbalance between supply and demand, some requests may be rejected or workers may keep idle for a long time. Fortunately, sharing tasks and workers across multiple platforms in a collaborative manner can diminish the negative effects of non-uniform distribution. In this paper, we propose a Multi-platform Route Planning in ridesharing problem (MPRP), which integrates route planning and task assignment in spatial crowdsourcing. We study how to improve the total revenue of platforms through coordination with multiple ridesharing platforms. To solve the problem, we propose the Greedy Multi-platform Route Planning algorithm (G-MPRP), which extends the dynamic programming insertion operation and assigns workers in a greedy manner. To overcome the shortcoming of G-MPRP, we further propose the Pack-based Multi-platform Route Planning algorithm (P-MPRP), which packs requests through a ranking-based function. Extensive experimental results shows the proposed algorithms can improve the revenues significantly.