[Paper Review] Large Scale Real-time Ridesharing with Service Guarantee on Road Networks

Yan Huang, Ruoming Jin+2 more|arXiv (Cornell University)|Feb 27, 2013

Transportation and Mobility Innovations13 references4 citations

TL;DR

This paper proposes a kinetic tree algorithm for large-scale real-time ridesharing on road networks, enabling dynamic matching of trip requests to servers while guaranteeing waiting and service time constraints. The method outperforms branch-and-bound and mixed-integer programming in response time on a large Shanghai taxi dataset, achieving faster real-time scheduling with strong service guarantees.

ABSTRACT

The mean occupancy rates of personal vehicle trips in the United States is only 1.6 persons per vehicle mile. Urban traffic gridlock is a familiar scene. Ridesharing has the potential to solve many environmental, congestion, and energy problems. In this paper, we introduce the problem of large scale real-time ridesharing with service guarantee on road networks. Servers and trip requests are dynamically matched while waiting time and service time constraints of trips are satisfied. We first propose two basic algorithms: a branch-and-bound algorithm and an integer programing algorithm. However, these algorithm structures do not adapt well to the dynamic nature of the ridesharing problem. Thus, we then propose a kinetic tree algorithm capable of better scheduling dynamic requests and adjusting routes on-the-fly. We perform experiments on a large real taxi dataset from Shanghai. The results show that the kinetic tree algorithm is faster than other algorithms in response time.

Motivation & Objective

• To address the challenge of large-scale, real-time ridesharing with strict service guarantees on road networks.
• To develop a dynamic matching algorithm that handles continuously arriving trip requests while satisfying waiting time and detour constraints.
• To minimize server travel time and maximize request acceptance in real-time, scalable systems.
• To overcome limitations of traditional branch-and-bound and mixed-integer programming approaches in dynamic, large-scale environments.
• To enable practical deployment in urban taxi services, private carpooling, and on-demand mobility systems.

Proposed method

• Proposes a kinetic tree data structure that maintains and updates candidate server assignments for incoming trip requests in real time.
• Uses a branch-and-bound algorithm and mixed-integer programming as baseline methods for comparison, though they are not scalable to dynamic, large-scale settings.
• Designs a kinetic tree that efficiently tracks the evolving state of candidate matches and updates them only when necessary, reducing computational overhead.
• Employs shortest path computation and detour cost estimation using Dijkstra’s algorithm to evaluate service constraints (1+ε)×d(s,e) for each request.
• Integrates real-time request evaluation by checking whether a server can serve a new request without violating existing waiting and detour constraints.
• Optimizes the kinetic tree with pruning and event-driven updates to maintain efficiency during high-volume request streams.

Experimental results

Research questions

• RQ1 How can real-time ridesharing be efficiently scaled to large urban road networks with thousands of taxis and tens of thousands of daily trips?
• RQ2 What algorithmic approach can dynamically match trip requests to servers while guaranteeing waiting time and detour constraints?
• RQ3 How does the performance of a kinetic tree compare to traditional exact methods like branch-and-bound and mixed-integer programming in real-time settings?
• RQ4 Can the kinetic tree maintain low response time and high request acceptance rates under dynamic, high-volume request workloads?
• RQ5 To what extent can the proposed method outperform heuristic-based approaches in terms of service guarantees and scalability?

Key findings

• The kinetic tree algorithm significantly reduces response time compared to branch-and-bound and mixed-integer programming on a large-scale Shanghai taxi dataset.
• The kinetic tree achieves faster real-time response by minimizing unnecessary recomputation through event-driven updates and pruning.
• The algorithm successfully maintains service guarantees, including maximum waiting time and detour constraints, across all matched requests.
• The method scales effectively to large urban road networks with up to 40,000 taxis and 400,000 daily trips, demonstrating practical viability.
• The kinetic tree outperforms baseline methods in both speed and scalability, making it suitable for real-world deployment in on-demand mobility services.
• The results confirm that the kinetic tree is a viable alternative to exact methods in dynamic, large-scale ridesharing systems where real-time performance is critical.