A SITING URBAN TAXI STATIONS MODEL BASED ON SPATIAL-TEMPORAL ORIGIN-DESTINATION DATA

To accommodate travel demands of residents and alleviate environmental impact of taxis, it is imperative to provide a plausible scheme for the construction of taxi stations in city. Moreover, it is more valuable to have stations in hotspots where there is more travel demand. The key to site taxi stations is stations location selection and parking spaces allocation in each station. The spatial-temporal origin-destination (OD) data are adopted, which can accurately locate passenger pick-up and drop-off locations, providing crucial data support for the establishment of taxi stations. Firstly, we extracted hotspot areas by calculating the number of trips and their economic value of each area to obtain a set of candidate areas. Given that site siting problem has the property of NP-Hard, we employ a heuristic genetic algorithm to derive an approximately feasible solution. We define the objective function by maximizing the total profit of stations and use queuing theory to determine the number of parking spaces in each station. Finally, the objective function is optimized through genetic operation in a finite number of iterations, with the optimal result being picked after several repetitions of the experiment. Our approach yielded promising performance on the Shijiazhuang dataset. We also analyzed the effect of the distances between demand area and station area, waiting times on station results. Our model was compared with a loss-based queueing theory model, and the results showed that our model performed significantly better in terms of objective function values. We also propose an improved adaptive genetic algorithm (AGA) to speed up the evolution of population. The findings indicate that our proposed methodology is well qualified to offer recommendations and reference for transport strategies and government decision-making.