The scheduling methods with different demand priorities for shared autonomous vehicle system in hybrid demands mode considering dynamic travel time

Abstract

The shared autonomous vehicle (SAV) is an emerging intelligent transportation mode driven by artificial intelligence, and the vehicle scheduling method is a crucial technology for SAVs. However, researches on the scheduling method of hybrid demand in SAVs lack consideration of dynamic travel time and an analysis of the impact of demand characteristics on system performances. According to the distinct submission and processing approaches of reservation and real-time demands, this paper proposes two hybrid demands scheduling methods for SAV system: the scheduling method with no demand priority (NDP) and the scheduling method that prioritizing short-term reservation demands (PSRD). The NDP method processes the short-term reservation and real-time demands together. The PSRD method processes the short-term reservation demands first, then inserts other requests into the arrangements of reservation demands. A method for calculating dynamic travel time using the ant colony algorithm is proposed. This method aims to identify the optimal path according to the future road network conditions. The results indicate that the PSRD method can give full play to the advantage of knowing the travel information of reservation demand in advance and has good system performance and running efficiency, especially when the number of reservation demands is large. This forms the foundation for advancing research in the SAV traffic systems.