Recommend a profitable cruising route for taxi drivers

Every time after the taxi drivers drop off their previous passengers, they have to decide how to search for the next passengers. It may lead to poor performance if the taxi drivers cruise according to their own experience, especially for the freshmen. Therefore, it is valuable to recommend a profitable cruising route for the taxi drivers in order to increase their income and reduce waste in fuel. In this paper, we propose to use a system of linear equations to calculate the score of each road segment based on a large-scale real-world GPS data set. The score of each road segment consists of 1) the total income of the road segment and 2) the attractiveness of the drop-off location with respect to the next pick-up. Then, we get the profitable cruising route based on the score of each road segment using skyline computation. We build our system using historical data set generated by 12,000 taxis of Beijing in November 2012. Finally, we demonstrate that taxi driver who follows our recommendation can enhance their income compared to the ground truth.