Rafal Kucharski Lab

2026

arXiv

Regulating Ride-Sourcing Markets: Can Minimum Wage Regulation Protect Drivers Without Disrupting the Market?

Ghasemi, Farnoud, Ruijter, Arjan, Kucharski, Rafal, and Cats, Oded

arXiv preprint arXiv:2509.11845 2026

Abs HTML

Ride-sourcing platforms such as Uber and Lyft are prime examples of the gig economy, recruiting drivers as independent contractors, thereby avoiding legal and fiscal obligations. Although platforms offer flexibility in choosing work shifts and areas, many drivers experience low income and poor working conditions, leading to widespread strikes and protests. Minimum wage regulation is adopted to improve drivers welfare. However, the impacts of this regulation on drivers as well as on travelers and platforms, remain largely unknown. While ride-sourcing platforms do not disclose the relevant data, state-of-the-art models fail to explain the effects of minimum wage regulation on market dynamics. In this study, we assess the effectiveness and implications of minimum wage regulation in ride-sourcing markets while simulating the detailed dynamics of ride-sourcing markets under varying regulation intensities, both with and without the so-called platform lockout strategy. Our findings reveal that minimum wage regulation impacts substantially drivers income, and may lead to higher fares for travelers and threaten platforms survival. When platforms adopt a lockout strategy, their profitability significantly improves and drivers earn more, although many others lose their jobs, and service level for travelers consequently declines.

2025

EJOR

Balancing profit and traveller acceptance in ride-pooling personalised fares

Bujak, Michał, and Kucharski, Rafał

European Journal of Operational Research 2025

Abs HTML

In a ride-pooling system, travellers experience discomfort associated with a detour and a longer travel time, which is compensated with a sharing discount. Most studies assume homogeneous travellers that receive either a flat discount or, in rare cases, a proportional to the inconvenience. This simplified approach offers inaccurate results and leads to an underperforming service when tested against diverse and natural human behaviour. We improve the standard approach on two bases. First, we propose a stochastic setting, where we leverage the population distribution of behavioural traits to determine the acceptance probability. Second, we personalise fares. Each traveller receives a sharing discount based on their contribution to the system such that the operator maximises his expected profitability. In the study, we rigorously prove that the discount optimisation problem can be decomposed. We optimise discounts at a ride level to claim the system optimum. An operator, when proposing fares, encounters two counteracting effects. Low fares increase realisation probability while high fares improve profit from a realised ride. In the personalised discount optimisation, we seek the golden mean. Travellers, who are well-aligned and experience minimal discomfort of sharing, are offered higher fares than those who require more incentive to join the service. Unlike in previous methods, our approach naturally balances the travellers satisfaction and the profit maximisation. With an experiment set in NYC, we show that this leads to significant improvements over the flat discount baseline: the mileage is reduced by 4.5% and the operator generates more profit per mile (over 20% improvement)
TR:C

MoMaS: Two-sided Mobility Market Simulation Framework for Modeling Platform Growth Trajectories

Ghasemi, Farnoud, and Kucharski, Rafal

Transportation Research Part C: Emerging Technologies 2025

Abs HTML

Mobility platforms such as Uber and DiDi have been introduced in cities worldwide, each demonstrating varying degrees of success, employing diverse strategies, and exerting distinct impacts on urban mobility. We have observed various growth trajectories in two-sided mobility markets and understood the underlying mechanisms. However, to date, a realistic microscopic model of these markets including phenomena such as network effects has been missing. State-of-the-art methods well estimate the macroscopic equilibrium conditions in the market, but struggle to reproduce the individual human behavior behind and complex growth patterns sensitive to platform strategy and policies. To bridge this gap, we introduce the MoMaS (two-sided Mobility Market Simulation) framework to represent growth mechanism in two-sided mobility markets based on the realistic behavior adjustment of drivers and travelers reactive to platform strategy. In the proposed framework, traveler and driver agents learn the platform utility from multiple channels: their own experience, peers’ word-of-mouth, and the platform’s marketing, all-together constituting the agent’s perceived utility of the platform. Each of these channels is modeled and updated by our S-shaped learning model day-to-day which stabilizes, and at the same time, remains sensitive to the system changes. The platform can simulate any strategy on five levers: trip fare, commission rate, discount rate, incentive rate, and marketing. While detailed empirical data and actual strategies for platform growth remain largely unknown, MoMaS allows to reproduce series of plausible growth trajectories that were previously unattainable. The framework facilitates the modeling of individual-level behaviors such as reluctance, neutrality, and loyalty, alongside aggregate-level dynamics like critical mass, bandwagon effects, and both positive and negative cross-side network effects. We illustrate the capabilities of MoMaS through an extensive set of real-world experiments. Our results demonstrate that once the platform acquires critical mass, it triggers a significant positive cross-side network effect, accelerating growth. However, this can be reversed if a negative cross-side network effect is triggered, leading to the collapse of the platform. MoMaS is applicable for real-sized problems and available on public repository along with reproducible experiments.
arXiv

URB - Urban Routing Benchmark for RL-equipped Connected Autonomous Vehicles

Akman, Ahmet Onur, Psarou, Anastasia, Hoffmann, Michał, Gorczyca, Łukasz, Kowalski, Łukasz, Gora, Paweł, Jamróz, Grzegorz, and Kucharski, Rafał

arXiv preprint arXiv:2505.17734 2025

Abs HTML

Connected Autonomous Vehicles (CAVs) promise to reduce congestion in future urban networks, potentially by optimizing their routing decisions. Unlike for human drivers, these decisions can be made with collective, data-driven policies, developed by machine learning algorithms. Reinforcement learning (RL) can facilitate the development of such collective routing strategies, yet standardized and realistic benchmarks are missing. To that end, we present \our: Urban Routing Benchmark for RL-equipped Connected Autonomous Vehicles. \our is a comprehensive benchmarking environment that unifies evaluation across 29 real-world traffic networks paired with realistic demand patterns. \our comes with a catalog of predefined tasks, four state-of-the-art multi-agent RL (MARL) algorithm implementations, three baseline methods, domain-specific performance metrics, and a modular configuration scheme. Our results suggest that, despite the lengthy and costly training, state-of-the-art MARL algorithms rarely outperformed humans. Experimental results reported in this paper initiate the first leaderboard for MARL in large-scale urban routing optimization and reveal that current approaches struggle to scale, emphasizing the urgent need for advancements in this domain.
SoftwareX

RouteRL: Multi-agent reinforcement learning framework for urban route choice with autonomous vehicles

Akman, Ahmet Onur, Psarou, Anastasia, Gorczyca, Łukasz, Varga, Zoltán György, Jamróz, Grzegorz, and Kucharski, Rafał

SoftwareX 2025

Abs HTML

RouteRL is a novel framework that integrates multi-agent reinforcement learning (MARL) with a microscopic traffic simulation for the development of efficient collective route choice strategies for autonomous vehicles (AVs). The proposed framework models the daily urban route choices of driver agents of two types: human drivers, emulated using behavioral route choice models, and AVs, modeled as MARL agents optimizing their policies for a predefined objective. RouteRL aims to advance research in MARL, transport modeling, and human–AI interaction for transportation applications. This study presents a technical report on RouteRL, outlines its potential research contributions, and showcases its impact via illustrative examples. Initial findings show that human travel times could increase following the introduction of AVs in urban environments, highlighting the importance of future studies to support efficient urban mobility.
Sci. Rep.

Social implications of coexistence of CAVs and human drivers in the context of route choice

Jamróz, Grzegorz, Akman, Ahmet Onur, Psarou, Anastasia, Varga, Zoltán György, and Kucharski, Rafał

Scientific Reports 2025

Abs HTML

Suppose in a stable urban traffic system populated only by human driven vehicles (HDVs), a given proportion (e.g. 10 %) is replaced by a fleet of Connected and Autonomous Vehicles (CAVs), which share information and pursue a collective goal. Suppose these vehicles are centrally coordinated and differ from HDVs only by their collective capacities allowing them to make more efficient routing decisions before the travel on a given day begins. Suppose there is a choice between two routes and every day each driver makes a decision which route to take. Human drivers maximize their utility. CAVs might optimize different goals, such as the total travel time of the fleet. We show that in this plausible futuristic setting, the strategy CAVs are allowed to adopt may result in human drivers either benefitting or being systematically disadvantaged and urban networks becoming more or less optimal. Consequently, some regulatory measures might become indispensable.
SNAM

Optimising network efficiency in an epidemic scenario

Proszewska, Magdalena, Bujak, Michal, Kucharski, Rafal, Tabor, Jacek, and Smieja, Marek

Social Network Analysis and Mining 2025

Abs HTML

Efficiency of a system network often relies on high connectivity. However, strongly connected networks are vulnerable in a case of a spreading virus. In our study, we propose a clustering method which balances the two opposing factors: maintains a high system efficiency yet minimises the spreading potential. Our Deep Epidemic Efficiency Network (DEEN) model leverages Graph Convolutional Neural Networks and a novel loss function. In an unsupervised setting, we seek a partition that maximises the system utility while restraining the transmission rate to a desired level. We show that proposed method successfully solves three real-life problems: ride-pooling service in New York City, economic exchange between regions in Poland, and information sharing via peer-to-peer network. In particular, by dividing 150 New York taxi travellers into four groups our method increases epidemic threshold more than twofold at the cost of reducing utility only by 13%. The model can be instrumental in future pandemic outbreaks when we need to balance between efficiency and potential spread of a virus.
J. public transp.

SimFLEX: A methodology for comparative analysis of urban areas for implementing new on-demand feeder bus services

Vasiutina, Hanna, Shulika, Olha, Bujak, Michał, Ghasemi, Farnoud, and Kucharski, Rafał

Journal of Public Transportation 2025

Abs HTML

On-demand feeder bus services present an innovative solution to urban mobility challenges, yet their success depends on a thorough assessment and strategic planning. Despite their potential, a systematic methodology for selecting suitable service areas remains underdeveloped. Simulation Framework for Feeder Location Evaluation (SimFLEX) utilizes spatial, demographic, and transportation-specific characteristics to conduct simulations at a microscopic level and compute various key performance indicators (KPIs), including service attractiveness, waiting time reduction, and added value. To address the stochastic nature of demand and uncertainty embedded in mode choice, we leverage Monte Carlo analysis, capturing variability across simulated scenarios. Our framework integrates several methods for a complete assessment of feeder potential: microscopic demand generation, creation of shared rides, public transport modeling, and iterative traveler learning and stabilization approach. Once the system stabilizes, KPIs are computed for comparative and sensitivity analyzes. As a showcase for our method, we apply SimFLEX to compare two remote districts in Krakow, Poland – Bronowice and Skotniki – designated as candidates for service deployment. Despite similar urban characteristics, our analysis revealed notable differences in KPIs between the analyzed areas: Skotniki exhibited higher service attractiveness (around 30 %) and added value (up to 7 %), whereas Bronowice showed greater potential for reducing waiting times (around 77 %). To assess the robustness of SimFLEX outputs under uncertain behavioral assumptions, we conducted a sensitivity analysis across a range of alternative-specific constants. The results consistently confirmed Skotniki as the more suitable candidate for feeder service implementation, demonstrating that this conclusion holds despite variations in a key preference parameter. The SimFLEX framework can be instrumental for decision makers to estimate feeder potential for a designated area. The model’s flexibility and modular characteristics make it a versatile tool for policymakers and urban planners to enhance urban mobility.
T&T

Can we start sharing our rides again? The post-pandemic ride-pooling market

Shulika, Olha, and Kucharski, Rafal

Transport and Telecommunication 2025

Abs HTML

Before the pandemic, ride-pooling was a promising mode of urban mobility, marked by increasing service providers and increasing traveller adoption, critical to its efficiency and sustainability. However, the COVID-19 pandemic caused significant disruption, with services suspended, business models altered, and reduced traveller confidence. In the post-pandemic era, understanding the future of ride-pooling is crucial. This article reviews the market through literature, pooling availability, and traveller behaviour studies. We find that the core elements of the ride-pooling model remain intact, with potential to appeal to travellers, drivers, platforms, and policymakers. Changes in travel behaviour due to the pandemic appear temporary, with a high willingness to share rides and reduce costs. However, whether ride-pooling can regain its growth remains uncertain. Despite unprecedented start-up activity, the financial prospects are unclear, posing challenges to its resurgence as a sustainable mobility solution.
arXiv

Impact of collective behaviors of autonomous vehicles on urban traffic dynamics: A multi-agent reinforcement learning approach

Akman, Ahmet Onur, Psarou, Anastasia, Varga, Zoltán György, Jamróz, Grzegorz, and Kucharski, Rafał

arXiv preprint arXiv:2509.22216 2025

Abs HTML

This study examines the potential impact of reinforcement learning (RL)-enabled autonomous vehicles (AV) on urban traffic flow in a mixed traffic environment. We focus on a simplified day-to-day route choice problem in a multi-agent setting. We consider a city network where human drivers travel through their chosen routes to reach their destinations in minimum travel time. Then, we convert one-third of the population into AVs, which are RL agents employing Deep Q-learning algorithm. We define a set of optimization targets, or as we call them behaviors, namely selfish, collaborative, competitive, social, altruistic, and malicious. We impose a selected behavior on AVs through their rewards. We run our simulations using our in-house developed RL framework PARCOUR. Our simulations reveal that AVs optimize their travel times by up to 5%, with varying impacts on human drivers’ travel times depending on the AV behavior. In all cases where AVs adopt a self-serving behavior, they achieve shorter travel times than human drivers. Our findings highlight the complexity differences in learning tasks of each target behavior. We demonstrate that the multi-agent RL setting is applicable for collective routing on traffic networks, though their impact on coexisting parties greatly varies with the behaviors adopted.
arXiv

Autonomous vehicles need social awareness to find optima in multi-agent reinforcement learning routing games

Psarou, Anastasia, Gorczyca, Łukasz, Gaweł, Dominik, and Kucharski, Rafał

arXiv preprint arXiv:2510.11410 2025

Abs HTML

Previous work has shown that when multiple selfish Autonomous Vehicles (AVs) are introduced to future cities and start learning optimal routing strategies using Multi-Agent Reinforcement Learning (MARL), they may destabilize traffic systems, as they would require a significant amount of time to converge to the optimal solution, equivalent to years of real-world commuting. We demonstrate that moving beyond the selfish component in the reward significantly relieves this issue. If each AV, apart from minimizing its own travel time, aims to reduce its impact on the system, this will be beneficial not only for the system-wide performance but also for each individual player in this routing game. By introducing an intrinsic reward signal based on the marginal cost matrix, we significantly reduce training time and achieve convergence more reliably. Marginal cost quantifies the impact of each individual action (route-choice) on the system (total travel time). Including it as one of the components of the reward can reduce the degree of non-stationarity by aligning agents’ objectives. Notably, the proposed counterfactual formulation preserves the system’s equilibria and avoids oscillations. Our experiments show that training MARL algorithms with our novel reward formulation enables the agents to converge to the optimal solution, whereas the baseline algorithms fail to do so. We show these effects in both a toy network and the real-world network of Saint-Arnoult. Our results optimistically indicate that social awareness (i.e., including marginal costs in routing decisions) improves both the system-wide and individual performance of future urban systems with AVs.
arXiv

Equilibria in routing games with connected autonomous vehicles will not be strong, as exclusive clubs may form

Kucharski, Rafał, Psarou, Anastasia, and Descormier, Natello

arXiv preprint arXiv:2510.12862 2025

Abs HTML

User Equilibrium is the standard representation of the so-called routing game in which drivers adjust their route choices to arrive at their destinations as fast as possible. Asking whether this Equilibrium is strong or not was meaningless for human drivers who did not form coalitions due to technical and behavioral constraints. This is no longer the case for connected autonomous vehicles (CAVs), which will be able to communicate and collaborate to jointly form routing coalitions. We demonstrate this for the first time on a carefully designed toy-network example, where a ‘club‘ of three autonomous vehicles jointly decides to deviate from the user equilibrium and benefit (arrive faster). The formation of such a club has negative consequences for other users, who are not invited to join it and now travel longer, and for the system, making it suboptimal and disequilibrated, which triggers adaptation dynamics. This discovery has profound implications for the future of our cities. We demonstrate that, if not prevented, CAV operators may intentionally disequilibrate traffic systems from their classic Nash equilibria, benefiting their own users and imposing costs on others. These findings suggest the possible emergence of an exclusive CAV elite, from which human-driven vehicles and non-coalition members may be excluded, potentially leading to systematically longer travel times for those outside the coalition, which would be harmful for the equity of public road networks.
arXiv

Wardropian Cycles make traffic assignment both optimal and fair by eliminating price-of-anarchy with Cyclical User Equilibrium for compliant connected autonomous vehicles

Hoffmann, Michał, Bujak, Michał, Jamróz, Grzegorz, and Kucharski, Rafał

arXiv preprint arXiv:2507.19675 2025

Abs HTML

Connected and Autonomous Vehicles (CAVs) open the possibility for centralised routing with full compliance, making System Optimal traffic assignment attainable. However, as System Optimum makes some drivers better off than others, voluntary acceptance seems dubious. To overcome this issue, we propose a new concept of Wardropian cycles, which, in contrast to previous utopian visions, makes the assignment fair on top of being optimal, which amounts to satisfaction of both Wardrop’s principles. Such cycles, represented as sequences of permutations to the daily assignment matrices, always exist and equalise, after a limited number of days, average travel times among travellers (like in User Equilibrium) while preserving everyday optimality of path flows (like in System Optimum). We propose exact methods to compute such cycles and reduce their length and within-cycle inconvenience to the users. As identification of optimal cycles turns out to be NP-hard in many aspects, we introduce a greedy heuristic efficiently approximating the optimal solution. Finally, we introduce and discuss a new paradigm of Cyclical User Equilibrium, which ensures stability of optimal Wardropian Cycles under unilateral deviations. We complement our theoretical study with large-scale simulations. In Barcelona, 670 vehicle-hours of Price-of-Anarchy are eliminated using cycles with a median length of 11 days-though 5% of cycles exceed 90 days. However, in Berlin, just five days of applying the greedy assignment rule significantly reduces initial inequity. In Barcelona, Anaheim, and Sioux Falls, less than 7% of the initial inequity remains after 10 days, demonstrating the effectiveness of this approach in improving traffic performance with more ubiquitous social acceptability.
arXiv

Autonomous Vehicles Using Multi-Agent Reinforcement Learning for Routing Decisions Can Harm Urban Traffic

Psarou, Anastasia, Akman, Ahmet Onur, Gorczyca, Łukasz, Hoffmann, Michał, Varga, Zoltán György, Jamróz, Grzegorz, and Kucharski, Rafał

arXiv preprint arXiv:2502.13188 2025

Abs HTML

Autonomous vehicles (AVs) using Multi-Agent Reinforcement Learning (MARL) for simultaneous route optimization may destabilize traffic environments, with human drivers possibly experiencing longer travel times. We study this interaction by simulating human drivers and AVs. Our experiments with standard MARL algorithms reveal that, even in trivial cases, policies often fail to converge to an optimal solution or require long training periods. The problem is amplified by the fact that we cannot rely entirely on simulated training, as there are no accurate models of human routing behavior. At the same time, real-world training in cities risks destabilizing urban traffic systems, increasing externalities, such as CO2 emissions, and introducing non-stationarity as human drivers adapt unpredictably to AV behaviors. Centralization can improve convergence in some cases, however, it raises privacy concerns for the travelers’ destination data. In this position paper, we argue that future research must prioritize realistic benchmarks, cautious deployment strategies, and tools for monitoring and regulating AV routing behaviors to ensure sustainable and equitable urban mobility systems.
TR:A

The implications of drivers’ ride acceptance decisions on the operations of ride-sourcing platforms

Ashkrof, Peyman, Ghasemi, Farnoud, Kucharski, Rafał, Almeida Correia, Gonçalo Homem, Cats, Oded, and Arem, Bart

Transportation Research Part A: Policy and Practice 2025

Abs HTML

As a two-sided digital platform, ride-sourcing has disruptively penetrated the mobility market. Ride-sourcing companies provide door-to-door transport services by connecting passengers with independent service suppliers labelled as “driver-partners”. Once a passenger submits a ride request, the platform attempts to match the request with a nearby available driver. Drivers have the freedom to accept or decline ride requests. The consequences of this decision, which is made at the operation level, have remained largely unknown in the literature. Using agent-based simulation modelling on the realistic case study of the city of Amsterdam, the Netherlands, we study the impacts of drivers’ ride acceptance behaviour, estimated from unique empirical data, on the ride-sourcing system where the platform applies regular and surge pricing strategies, and riders may revoke their requests and reject the received offers. Furthermore, we delve into the implications of various supply–demand intensities, a centralised fleet (i.e., mandatory acceptance on each ride request) versus a decentralised fleet (i.e., ride acceptance decision by each driver), ride acceptance rates, and surge pricing settings. We find that the ride acceptance decision of ride-sourcing drivers has far-reaching consequences for system performance in terms of passengers’ waiting time, driver’s revenue, operating costs, and profit, all of which are highly dependent on the ratio between demand and supply. As the system undergoes a transition from undersupplied (i.e., real-time demand locally exceeds available drivers) to balanced and then oversupplied state (i.e., more available drivers than real-time demand), ride acceptance decisions result in higher income inequality. A high acceptance rate among drivers may lead to more rides, but it does not necessarily increase their profit. Surge pricing is found to be asymmetrically in favour of all the parties despite adverse effects on the demand side due to higher trip fare. This study offers insights into both the aggregated and disaggregated levels of ride-sourcing system operations and outlines a series of transport policy and practice implications in cities that offer such ride-sourcing systems.

2024

ITSC

Reinforcement Learning Approach for Improving Platform Performance in Two-Sided Mobility Markets

Ghasemi, Farnoud, Tabatabaei, Seyed Hassan, Ghanadbashi, Saeedeh, Kucharski, Rafał, and Golpayegani, Fatemeh

In 2024 IEEE 27th International Conference on Intelligent Transportation Systems (ITSC) 2024

Abs HTML

Two-sided mobility markets, with platforms like Uber and Lyft, are complex systems by nature due to intricate, non-linear interactions between the platform and the involved parties including travelers and drivers. These interactions give rise to phenomena underlying market evolution, mainly cross-side network effects. Currently, such platforms rely on rule-based (RB) strategies with a constant commission rate to grow and achieve sustainability in terms of market share and profitability. However, the constant commission rate significantly constrains the platform’s ability to leverage network effects, leading to inefficient growth. In this study, a Reinforcement Learning-based (RLB) strategy is proposed to improve the platform performance through strategic levers. We employ a Deep Q-Network (DQN) within an agent-based framework, enabling the platform to adjust the commission rate on a day-to-day basis while learning the complex, non-linear interactions in the market. The results show that the RL-based strategies successfully generate and control the essential cross-side network effects in the market enhancing the platform performance via dynamic commission rate. Our results indicate 12% improvement in the platform revenue with the RL-based strategy in comparison to the rule-based strategy without significantly compromising the platform market share which can essentially impact the platform’s viability in the long term.
npj

Hyper pooling private trips into high occupancy transit like attractive shared rides

Kucharski, Rafał, and Cats, Oded

npj Sustainable Mobility and Transport 2024

Abs HTML

The size of the solution space associated with the trip-matching problem has made the search for high-order ride-pooling prohibitive. We introduce hyper-pooled rides along with a method to identify them within urban demand patterns. Travellers of hyper-pooled rides walk to common pick-up points, travel with a shared vehicle along a sequence of stops and are dropped off at stops from which they walk to their destinations. While closely resembling classical mass transit, hyper-pooled rides are purely demand-driven, with itineraries (stop locations, sequences, timings) optimised for all co-travellers. For 2000 trips in Amsterdam the algorithm generated 40 hyper-pooled rides transporting 225 travellers. They would require 52.5 vehicle hours to travel solo, whereas in the hyper-pooled multi-stop rides, it is reduced sixfold to 9 vehicle hours only. This efficiency gain is made possible by achieving an average occupancy of 5.8 (and a maximum of 14) while remaining attractive for all co-travellers.
Transportation

Ride-pooling service assessment with heterogeneous travellers in non-deterministic setting

Bujak, Michal, and Kucharski, Rafal

Transportation 2024

Abs HTML

Ride-pooling remains a promising emerging mode with a potential to contribute towards urban sustainability and emission reductions. Recent studies revealed complexity and diversity among travellers’ ride-pooling attitudes. So far, ride-poling analyses assumed homogeneity of ride-pooling travellers. This, as we demonstrate, leads to a false assessment of ride-pooling system performance. We experiment with an actual NYC demand from 2016 and classify travellers into four groups of various ride-pooling behaviours (value of time and penalty for sharing), as reported in the recent SP study from Netherlands. We replicate their behavioural characteristics, according to the population distribution, to obtain meaningful performance estimations. Results vary significantly from the homogeneous benchmark: mileage savings were lower, while the utility gains for travellers were greater. Observing performance of heterogeneous travellers, we find that those with a low value of time are most beneficial travellers in the pooling system, while those with an average penalty for sharing benefit the most. Notably, despite the highly variable travellers’ behaviour, the confidence intervals for the key performance indicators are reasonably narrow and system-wide performance remains predictable. Our results show that the incorrect assumption of homogeneous traits leads to a high dissatisfaction of 18.5% and a cancellation rate of 36%. Such findings shed a new light on the expected performance of large scale ride-pooling systems.
JoTG

Spatiotemporal variability of ride-pooling potential – Half a year New York City experiment

Shulika, Olha, Bujak, Michal, Ghasemi, Farnoud, and Kucharski, Rafal

Journal of Transport Geography 2024

Abs HTML

Ride-pooling systems, despite being an appealing urban mobility mode, still struggle to gain momentum. While we know the significance of critical mass in reaching system sustainability, less is known about the spatiotemporal patterns of system performance. Here, we use 1.5 million NYC taxi trips (sampled over a six-month period) and experiment to understand how well they could be served with pooled services. We use an offline utility-driven ride-pooling algorithm and observe the pooling potential with six performance indicators: mileage reductions, travellers’ utility gains, share of pooled rides, occupancy, detours, and potential fleet reduction. We report distributions and temporal profiles of about 35 thousand experiments covering weekdays, weekends, evenings, mornings, and nights. We report complex spatial patterns, with gains concentrated in the core of the network and costs concentrated on the peripheries. The greatest potential shifts from the North in the morning to the Central and South in the afternoon. Offering pooled rides at the fare 32% lower than private ride-hailing seems to be sufficient to attract pooling yet dynamically adjusting it to the demand level and spatial pattern may be efficient. The patterns observed in NYC were replicated on smaller datasets in Chicago and Washington, DC, the occupancy grows with the demand with similar trends.
AAMAS

Modelling the Rise and Fall of Two-sided Markets

Ghasemi, Farnoud, and Kucharski, Rafal

In Proceedings of the 23rd International Conference on Autonomous Agents and Multiagent Systems 2024

Abs HTML

Two-sided markets disrupted our economies, reshaping markets as diverse as tourism (airbnb), mobility (Uber) and food deliveries (UberEats). New market leaders arose leveraging on platform-based business model, questioning well-established paradigms. The underlying processes behind their growth are non-trivial, inherently microscopic, and leverage on complex human interactions. Platforms need to reach critical mass of both supply and demand to trigger the so-called cross-sided network effects. To this end, platforms adopt a variety of strategies to first create the market, then expand it and finally successfully compete with others. Such a complex social system with many non-linear interactions and learning processes calls for a dedicated modelling approach. State-of-the-art methods well estimate the macroscopic equilibrium conditions, but struggle to reproduce the complex growth patterns and individual human behaviour behind. To bridge this gap, we propose the microscopic S-shaped learning model where agents build their perception on the new service with time, affected by both endogenous (service quality) and exogenous (marketing and word-of-mouth) factors cumulated from experiences. We illustrate it with the case of two-sided mobility platform (Uber), where the platform applies a series of marketing actions leading to rise and then fall on the market where 200 drivers serve 2000 travellers on the complex urban network of Amsterdam. Our model is the first to reproduce not only behaviourally sound, but also empirically observed growth trajectories, it remains sensitive to a variety of marketing strategies, allows reproducing the competition between platforms and is designed to be integrated with machine learning algorithms to identify the optimal market entry strategy.
arXiv

Optimizing Ride-Pooling Revenue: Pricing Strategies and Driver-Traveller Dynamics

Akhtar, Usman, Ghasemi, Farnoud, and Kucharski, Rafal

arXiv preprint arXiv:2403.13384 2024

Abs HTML

Ride-pooling, to gain momentum, needs to be attractive for all the parties involved. This includes also drivers, who are naturally reluctant to serve pooled rides. This can be controlled by the platform’s pricing strategy, which can stimulate drivers to serve pooled rides. Here, we propose an agent-based framework, where drivers serve rides that maximise their utility. We simulate a series of scenarios in Delft and compare three strategies. Our results show that drivers, when they maximize their profits, earn more than in both the solo-rides and only-pooled rides scenarios. This shows that serving pooled rides can be beneficial as well for drivers, yet typically not all pooled rides are attractive for drivers. The proposed framework may be further applied to propose discriminative pricing in which the full potential of ride-pooling is exploited, with benefits for the platform, travellers, and (which is novel here) to the drivers.

2023

SNAM

Network structures of urban ride-pooling problems and their properties

Bujak, Michal, and Kucharski, Rafal

Social Network Analysis and Mining 2023

Abs HTML

Travellers, when sharing their rides in a so-called ride-pooling system, form complex networks. Despite being the algorithmic backbone to the ride-pooling problems, the shareability graphs have not been explicitly analysed yet. Here, we formalise them, study their properties and analyse relations between topological properties and expected ride-pooling performance. We introduce and formalise two representations at the two crucial stages of pooling analysis. On the NYC dataset, we run two simulations with the link generation formulas. One is when we increase discount offered to the travellers for shared rides (our control variable) and observe the phase transition. In the second, we replicate the non-deterministic behaviour of travellers in ride-pooling. This way, we generate probabilistic, weighted networks. We observed a strong correlation between the topological properties of ride-pooling networks and the system performance. Introduced class of networks paves the road to applying the network science methods to a variety of ride-pooling problems, like virus spreading, optimal pricing or stability analysis.
arXiv

Ride Acceptance Behaviour Investigation of Ride-sourcing Drivers Through Agent-based Simulation

Ghasemi, Farnoud, Ashkrof, Peyman, and Kucharski, Rafal

arXiv preprint arXiv:2310.05588 2023

Abs HTML

Ride-sourcing platforms such as Uber and Lyft offer drivers (i.e., platform suppliers) considerable freedom of choice in multiple aspects. At the operational level, drivers can freely accept or decline trip requests that can significantly impact system performance in terms of travellers’ waiting time, drivers’ idle time and income. Despite the extensive research into the supply-side operations, the behavioural aspects, particularly drivers’ ride acceptance behaviour remains so far largely unknown. To this end, we reproduce the dynamics of a two-sided mobility platform on the road network of Delft using an agent-based simulator. Then, we implement a ride acceptance decision model enabling drivers to apply their acceptance strategies. Our findings reveal that drivers who follow the decision model, on average, earn higher income compared to drivers who randomly accept trip requests. The overall income equality between drivers with the acceptance decision is higher and travellers experience lower waiting time in this setting.
arXiv

Dynamics of the Ride-Sourcing Market: A Coevolutionary Model of Competition between Two-Sided Mobility Platforms

Ghasemi, Farnoud, Drabicki, Arkadiusz, and Kucharski, Rafał

arXiv preprint arXiv:2310.05543 2023

Abs HTML

There is a fierce competition between two-sided mobility platforms (e.g., Uber and Lyft) fueled by massive subsidies, yet the underlying dynamics and interactions between the competing plat-forms are largely unknown. These platforms rely on the cross-side network effects to grow, they need to attract agents from both sides to kick-off: travellers are needed for drivers and drivers are needed for travellers. We use our coevolutionary model featured by the S-shaped learning curves to simulate the day-to-day dynamics of the ride-sourcing market at the microscopic level. We run three scenarios to illustrate the possible equilibria in the market. Our results underline how the correlation inside the ride-sourcing nest of the agents choice set significantly affects the plat-forms’ market shares. While late entry to the market decreases the chance of platform success and possibly results in "winner-takes-all", heavy subsidies can keep the new platform in competition giving rise to "market sharing" regime.

2022

arXiv

Exploring Computational Complexity Of Ride-Pooling Problems

Akhtar, Usman, and Kucharski, Rafal

arXiv preprint arXiv:2208.02504 2022

Abs HTML

Ride-pooling is computationally challenging. The number of feasible rides grows with the number of travelers and the degree (capacity of the vehicle to perform a pooled ride) and quickly explodes to the sizes making the problem not solvable analytically. In practice, heuristics are applied to limit the number of searches, e.g., maximal detour and delay, or (like we use in this study) attractive rides (for which detour and delay are at least compensated with the discount). Nevertheless, the challenge to solve the ride-pooling remains strongly sensitive to the problem settings. Here, we explore it in more detail and provide an experimental underpinning to this open research problem. We trace how the size of the search space and computation time needed to solve the ride-pooling problem grows with the increasing demand and greater discounts offered for pooling. We run over 100 practical experiments in Amsterdam with 10-minute batches of trip requests up to 3600 trips per hour and trace how challenging it is to propose the solution to the pooling problem with our ExMAS algorithm. We observed strong, non-linear trends and identified the limits beyond which the problem exploded and our algorithm failed to compute. Notably, we found that the demand level (number of trip requests) is less critical than the discount. The search space grows exponentially and quickly reaches huge levels. However, beyond some level, the greater size of the ride-pooling problem does not translate into greater efficiency of pooling. Which opens the opportunity for further search space reductions.

2021

SciRep

Modelling virus spreading in ride-pooling networks

Kucharski, Rafał, Cats, Oded, and Sienkiewicz, Julian

Scientific Reports 2021

Abs HTML

Urban mobility needs alternative sustainable travel modes to keep our pandemic cities in motion. Ride-pooling, where a single vehicle is shared by more than one traveller, is not only appealing for mobility platforms and their travellers, but also for promoting the sustainability of urban mobility systems. Yet, the potential of ride-pooling rides to serve as a safe and effective alternative given the personal and public health risks considerations associated with the COVID-19 pandemic is hitherto unknown. To answer this, we combine epidemiological and behavioural shareability models to examine spreading among ride-pooling travellers, with an application for Amsterdam. Findings are at first sight devastating, with only few initially infected travellers needed to spread the virus to hundreds of ride-pooling users. Without intervention, ride-pooling system may substantially contribute to virus spreading. Notwithstanding, we identify an effective control measure allowing to halt the spreading before the outbreaks (at 50 instead of 800 infections) without sacrificing the efficiency achieved by pooling. Fixed matches among co-travellers disconnect the otherwise dense contact network, encapsulating the virus in small communities and preventing the outbreaks.