The Customer is Always Right: Customer-Centered Pooling for Ride-Hailing Systems

Paul Karaenke, Maximilian Schiffer, Stefan WaldherrarXiv 2021 · arXiv

Abstract

While ride-hailing services were initially expected to alleviate urban congestion, they have frequently exacerbated the problem due to low vehicle utilization and the inefficiencies of current pooling models. This paper introduces Customer-Centered Pooling (CCP), a mechanism that shifts the focus from fixed platform discounts to flexible, individual-based pricing. By incorporating a rider's "Value of Time" (VoT) into the fare calculation, CCP ensures that participating in a pool is a weakly dominant strategy for the customer. Our simulation results, based on New York City taxi data, demonstrate that CCP can achieve service rates of 99.97%, reduce total distance driven by up to 32%, and stabilize operator profits by eliminating the "sunk costs" of unmatched discounts.

Introduction

Ride-hailing platforms like Uber and Lyft face a persistent challenge: ride-pooling is the most effective way to reduce traffic and emissions, yet adoption remains hampered by a "chicken and egg" dilemma. For pooling to be efficient, a large participant pool is required; however, customers are often reluctant to join because current systems feel like a "lottery."

In the traditional Provider-Centered Pooling (PCP) model, platforms offer a fixed discount regardless of the actual detour or delay a rider experiences. This often results in "win-lose" scenarios: either the rider suffers an excessive delay for a negligible discount, or the provider loses money by offering a discount for a ride that ultimately fails to find a match. This paper addresses these issues of fairness and transparency by proposing a mechanism that aligns the interests of the rider, the operator, and the environment.

Methodology

The Customer-Centered Pooling (CCP) Mechanism

The core innovation of CCP is its departure from flat-rate discounting. Instead, it calculates fares based on the specific inconvenience—detours and wait times—experienced by each customer.

  1. Flexible Discounting: CCP utilizes the customer’s self-indicated Value of Time (VoT). A ride is only pooled if the total cost (the monetary fare plus the cost of the rider's time) is lower than the cost of a solo trip.
  2. Game-Theoretic Foundation: The authors prove that CCP is "individually rational" and a "weakly dominant strategy." In simpler terms, a customer is never worse off by opting into CCP, removing the psychological barrier of the "pooling lottery."
  3. Optimization and Fare Splitting: The researchers developed an Integer Program (IP) for ex-post fare splitting. This allows operators to maximize the number of riders receiving substantial savings, ensuring that the benefits of pooling are distributed equitably.

Simulation Environment

To test the mechanism, the authors built a Java-based simulation using the OSM2PO library for routing and a greedy assignment algorithm. The study utilized the New York Taxi and Limousine Commission (TLC) dataset, focusing on a representative sample of 9,388 trips in Manhattan (south of Central Park) during a peak morning hour. The simulation modeled a fleet of 1,000 to 3,000 vehicles, with customer VoT categorized into five distinct segments based on transportation economics literature.

Key Insights

The Profitability of Fairness

Counter-intuitively, prioritizing customer benefit does not diminish provider profits. Under the PCP model, providers often grant discounts upfront; if a match isn't found, the provider absorbs that loss. CCP ensures the operator only pays out a discount when a profitable, efficient match is actually made. This makes CCP more robust for operators, particularly in scenarios with fluctuating participation rates.

Pareto Dominance and Environmental Synergy

In many simulated scenarios, CCP "Pareto dominates" PCP, meaning it improves the outcome for at least one party without making anyone else worse off. Furthermore, because CCP only pools when it is truly efficient for the participants, it naturally optimizes for the shortest combined paths. This creates a direct link between customer satisfaction and sustainability, leading to significant reductions in CO2 emissions.

Results

  • Service Levels: CCP consistently outperformed both PCP and solo-ride baselines, reaching a near-perfect service level of 99.97% at high participation rates.
  • Distance and Emissions: The mechanism reduced the total distance driven in the system by up to 32%. In the Manhattan scenario, even a 5% reduction in distance equated to saving 0.16 metric tons of CO2 within a mere 30-minute window.
  • Economic Stability: CCP mechanisms showed stable profit growth. Unlike PCP, which can lead to losses at low participation rates due to uncompensated discounts, CCP ensures that every match contributes to the operator's bottom line.
  • Substantial Savings: Using the IP-based pricing model, the authors increased the share of customers receiving "substantial" savings (defined as over 15% of total costs) compared to standard methods like the Shapley value.

Limitations and Future Work

While the results are promising, the authors note several areas for further research. The current simulation relies on a greedy matching algorithm; more sophisticated "look-ahead" algorithms could potentially enhance performance further. Additionally, the model assumes customers are economically rational actors who accurately report their VoT. Future work should involve behavioral experiments to see how real-world users interact with "urgency menus" in a ride-hailing app. Finally, scaling the Integer Program for fare splitting to city-wide, real-time operations will require significant algorithmic optimization to handle tens of thousands of concurrent requests.

Summary updated

Abstract

Today's ride-hailing systems experienced significant growth and ride-pooling promises to allow for efficient and sustainable on-demand transportation. However, efficient ride-pooling requires a large pool of participating customers. To increase the customers' willingness for participation, we study a novel customer-centered pooling (CCP) mechanism, accounting for individual customers' pooling benefits. We study the benefit of this mechanism from a customer, fleet operator, and system perspective, and compare it to existing provider-centered pooling (PCP) mechanisms. We prove that it is individually rational and weakly dominant for a customer to participate in CCP, but not for PCP. We substantiate this analysis with complementary numerical studies, implementing a simulation environment based on real-world data that allows us to assess both mechanisms' benefit. To this end, we present results for both pooling mechanisms and show that pooling can benefit all stakeholders in on-demand transportation. Moreover, we analyze in which cases a CCP mechanism Pareto dominates a PCP mechanism and show that a mobility service operator would prefer CCP mechanisms over PCP mechanisms for all price segments. Simultaneously, CCP mechanisms reduce the overall distance driven in the system up to 32% compared to not pooling customers. Our results provide decision support for mobility service operators that want to implement and improve pooling mechanisms as they allow us to analyze the impact of a CCP and a PCP mechanism from a holistic perspective. Among others, we show that CCP mechanisms can lead to a win-win situation for operators and customers while simultaneously improving system performance and reducing emissions.
BibTeX
@misc{karaenke2021customer,
  author = {Paul Karaenke and Maximilian Schiffer and Stefan Waldherr},
  title = {{The Customer is Always Right: Customer-Centered Pooling for Ride-Hailing Systems}},
  year = {2021},
  eprint = {2107.01161},
  archivePrefix = {arXiv},
  url = {https://arxiv.org/abs/2107.01161},
}
Figure