TL;DR
This paper introduces a novel "Taxi-Sharing and Subway" (TSS) multimodal transit system designed to integrate taxi services with subways for a seamless door-to-door experience. By developing an optimization-based matching model, they demonstrate that TSS can significantly reduce taxi mileage and carbon emissions while offering cost savings to passengers, making it a sustainable and efficient urban mobility solution.
What Problem Does This Paper Address?
Urban areas worldwide grapple with severe traffic congestion and air pollution, largely driven by private vehicle use and inefficient public transportation. While subways are efficient for high-capacity transit, they often suffer from the "first/last mile" problem, requiring passengers to travel to and from stations. Taxis, while flexible, contribute to congestion and are often underutilized. Existing multimodal studies have largely overlooked the integration of taxi-sharing with subways as a comprehensive door-to-door service. This paper addresses the challenge of creating a cooperative relationship between taxis and subways, moving beyond the traditional feeder service model to a fully integrated "Mobility as a Service" (MaaS) offering that optimizes the entire journey.
Methods
The core of this research is a two-phase matching model for integrating taxi-sharing and subway services.
- Intermodal Transfer Station Optimization: The first phase identifies optimal subway stations for transfers. This involves mapping taxi trip origins and destinations to the nearest subway stations and then finding the shortest path between these "projective" subway stations within the subway network. This path represents an "alternative subway path," and the stations on this path are considered potential transfer points. The selection aims to minimize generalized travel costs, considering subway travel time, fares, taxi-sharing costs, and transfer penalties.
- Matching Model: The second phase employs an integer linear programming (ILP) model to optimize the matching of passengers into TSS trips. The objective function is to maximize the total taxi distance savings for the system. The model considers various trip subtypes (TST, TSW, WST), pick-up/drop-off sequences, and passenger preferences (modeled through generalized cost savings). It also incorporates constraints to ensure each passenger is matched at most once and that matched passengers realize a generalized cost saving compared to their original taxi trip.
Datasets
The study utilizes empirical taxi global positioning system (GPS) data from Beijing for a typical morning rush hour (June 18, 2014, 7:00-9:00 AM). This dataset includes 19,156 taxi trips with pick-up/drop-off times and locations. The subway network data for Beijing was also incorporated.
Key Insights and Ideas
- Beyond First/Last Mile: The research moves beyond viewing taxi-sharing solely as a feeder service, proposing a comprehensive door-to-door multimodal solution that optimizes the entire journey.
- Passenger-Centric Optimization: The model prioritizes passenger cost savings, ensuring that any shift to TSS transit results in a lower generalized travel cost for the individual.
- Trade-offs in Parameter Tuning: The study highlights that optimizing parameters like subway unit cost and transfer penalties involves trade-offs between match rate and distance savings, suggesting a need for balanced policy decisions.
- Attitudes Matter: Passenger attitudes, particularly towards subway travel time costs and transfer penalties, significantly influence the adoption and performance of TSS. Lower transfer penalties are crucial for higher adoption.
- Dynamic System Suitability: The optimization model is designed to be computationally efficient, capable of handling large-scale data and dynamic systems, making it practical for real-world implementation.
Key Results
- High Adoption Potential: Approximately 23.13% of potential taxi trips in Beijing could be successfully served by the proposed TSS multimodal transit system.
- Environmental Benefits: TSS transit can lead to a significant reduction in taxi mileage (20.17%) and carbon dioxide emissions (15.16%).
- Cost Savings: On average, passengers shifting to TSS can save approximately CNY 23.9 per trip, with a total travel cost-saving rate of 9.37%.
- Operational Efficiency: The model demonstrated computational efficiency, solving for over 19,000 trips in under 3 minutes.
- Parameter Sensitivity:
- A subway unit cost of 1.0 CNY/min was found to be the most appropriate for balancing match rate and distance savings in Beijing.
- Reducing the mode transfer penalty was found to be more effective than controlling subway costs for improving TSS performance.
- An optimal time similarity threshold of 15 minutes was identified to balance match rate and response time.
Limitations and Future Work
- Limited Validation: The study acknowledges that a real-world validation test of the TSS system is needed, as implementing and operating such a platform is complex and expensive.
- Simplified Cost Model: The generalized cost function currently considers limited primary factors and neglects distinctions between passengers, such as psychological costs of sharing or specific preferences.
- Scope of Application: The proposed approach is primarily suited for metropolitan areas with significant road congestion and air pollution. Its attractiveness may decrease in areas with inexpensive taxis and expensive subways.
- Matching Capacity: The current model is limited to matching two requests at a time; future work could explore matching larger groups.
- Equity Concerns: The study briefly touches upon potential equity issues, such as the reliance on smartphones and unequal cost-saving benefits within a matching group, which warrant further investigation.
Future work should focus on practical implementation, collecting real-world user feedback, incorporating more personalized demand factors, designing a better fare scheme, and addressing equity concerns to ensure the sustainable and inclusive development of TSS multimodal transit.
Summary updated