Feasibility and issues for establishing network-based carpooling scheme

Traffic issue has become one of the most concerns for citizens in metropolis. It is extremely hard to find an available vehicle in hot region during the peak hour. Due to the supply-demand contradiction, the authority is still seeking effective and economical carpooling solutions for \"Taxi taking dilemma\" without adding more vehicles. This paper mainly focuses on building a network-based carpooling scheme by analyzing the feasibility and related issues. Firstly, we provide some fundamental analysis for taxi carpooling based on the real taxis' GPS traces. By setting up a correspondence relationship between physical locations and the coordinates of map, a macro impression about the space and time distribution of get on points for Beijing taxis are illustrated. According to the prediction algorithm, both the original and estimated benefits of carpooling are shown quantitatively. Secondly, four different network-based sharing schemes are presented and compared. The conditions of successful route matching are qualitatively studied. By identifying the merits and weaknesses comprehensively, we propose a new hybrid carpooling scheme. The concerns relate with the interaction process and packet structuring are considered carefully. A novel perspective for distinguishing vehicle and pedestrian in software design is also pointed out.