An improved multi-objective method for the selection of driverless taxi site locations

To expedite the large-scale deployment of driverless taxis and advance the autonomous driving industry, research on the location of integrated parking and charging facilities for driverless taxis has emerged as a significant issue in urban traffic. This study employs a progressive “preliminary selection-screening-optimal selection” approach for site selection. First, the preliminary selection of parking sites is conducted by clustering various point-of-interest types. Subsequently, a multi-objective site selection model is developed to maximize the coverage of demand points, minimize construction costs, address the largest population demands, and minimize the distance between demand points and candidate sites. The non-dominated sorting genetic algorithm II (NSGA-II) is adopted to obtain several Pareto optimal solutions. The evaluation indexes are selected according to operators, users, and the public transport system to estimate the Pareto optimal solutions, and then the final location solution can be obtained. The calculation methods for several key parameters are improved during the modeling process. Location potential and location influence coefficient are selected to adjust the number of driverless taxi parking spaces. Additionally, isochrones drawn based on the actual road network and path planning represent the service range of candidate points. Meanwhile, distance based on actual road network rather than Euclidean distance is introduced to calculate the distance between candidate points. Finally, a case study shows that the method proposed in this study could reduce the total initial travel time to reach the demand points by 64%, which is independent of operational scheduling.