Empty-Car Routing in Ridesharing Systems

Braverman, Anton; Dai, J. G.; Liu, Xin; Lü, Ying

doi:10.1287/opre.2018.1822

Cited by 181 publications

(121 citation statements)

References 26 publications

Supporting

Mentioning

112

Contrasting

Order By: Relevance

“…Hence, the "size" of regions can be determined by the waiting-time limit set by the ride-sharing system. Our road network and region graph models are similar to those used in the literature; see e.g., [13,14,22]. All our modeling and analysis are based on the region graph G.…”

Section: Transportation Network and Region Graphmentioning

confidence: 99%

A Probabilistic Approach for Demand-Aware Ride-Sharing Optimization

Lin

Chen

et al. 2019

Proceedings of the Twentieth ACM International Symposium on Mobile Ad Hoc Networking and Computing

View full text Add to dashboard Cite

Ride-sharing is a modern urban-mobility paradigm with tremendous potential in reducing congestion and pollution. Demand-aware design is a promising avenue for addressing a critical challenge in ride-sharing systems, namely joint optimization of request-vehicle assignment and routing for a fleet of vehicles. In this paper, we develop a probabilistic demand-aware framework to tackle the challenge. We focus on maximizing the expected number of passenger pickups, given the probability distributions of future demands. The key idea of our approach is to assign requests to vehicles in a probabilistic manner. It differentiates our work from existing ones and allows us to explore a richer design space to tackle the requestvehicle assignment puzzle with a performance guarantee but still keeping the final solution practically implementable. The optimization problem is non-convex, combinatorial, and NP-hard in nature. As a key contribution, we explore the problem structure and propose an elegant approximation of the objective function to develop a dual-subgradient heuristic. We characterize a condition under which the heuristic generates a (1 − 1/e) approximation solution. Our solution is simple and scalable, amendable for practical implementation. Results of numerical experiments based on real-world traces in Manhattan show that, as compared to a conventional demand-oblivious scheme, our demand-aware solution improves the passenger pickups by up to 46%. The results also show that joint optimization at the fleet level leads to 19% more pickups than that by separate optimizations at individual vehicles.

show abstract

Section: Transportation Network and Region Graphmentioning

confidence: 99%

A Probabilistic Approach for Demand-Aware Ride-Sharing Optimization

Lin

Chen

et al. 2019

Proceedings of the Twentieth ACM International Symposium on Mobile Ad Hoc Networking and Computing

View full text Add to dashboard Cite

show abstract

“…In this paper, we assume that m c · n for some constant capacity c. A dispatching policy (or strategy) σ is a mapping from Ω × R to U ∪ {∅} such that at time t and a state X (t) , when a request r comes, σ assigns that request r to a potential driver at location u σ = σ(X (t) , r). 6 Here u σ = ∅ denotes that the policy σ rejects request r. We say σ successfully addresses…”

Section: Preliminariesmentioning

confidence: 99%

“…Research in rideshare platforms and similar allocation problems is an active area of research within multiple fields, including computer science, operations research and transportation engineering. Stateindependent policies were studied previously using theory from control and queuing systems [28,6,4]. Apart from using Markov-chain methods, many allocation and scheduling problems have been studied in the rideshare context using methods from combinatorial optimization and machine learning (e.g., [11,9,16]).…”

Section: Assumptions and Related Workmentioning

confidence: 99%

See 1 more Smart Citation

Mix and Match: Markov Chains and Mixing Times for Matching in Rideshare

Curry

Dickerson

Sankararaman

et al. 2019

Web and Internet Economics

View full text Add to dashboard Cite

Rideshare platforms such as Uber and Lyft dynamically dispatch drivers to match riders' requests. We model the dispatching process in rideshare as a Markov chain that takes into account the geographic mobility of both drivers and riders over time. Prior work explores dispatch policies in the limit of such Markov chains; we characterize when this limit assumption is valid, under a variety of natural dispatch policies. We give explicit bounds on convergence in general, and exact (including constants) convergence rates for special cases. Then, on simulated and real transit data, we show that our bounds characterize convergence rates-even when the necessary theoretical assumptions are relaxed. Additionally these policies compare well against a standard reinforcement learning algorithm which optimizes for profit without any convergence properties. IntroductionRideshare firms such as Uber, Lyft, and Didi Chuxing dynamically match riders to drivers via an online, digital platform. Riders request a driver through an online portal or mobile app; a driver is matched by the platform to a rider based on geographic proximity, driver preferences, pricing, and other factors. The rideshare driver then picks up the rider at her request location, transfers her to her destination, and reenters the platform to be matched again-albeit at a new geographic location. Part of the larger sharing economy, rideshare firms are increasingly competitive against traditional taxi services due to their ease of use, lower pricing, and immediacy of service [14].Matching riders to rideshare drivers is nontrivial. While the core process is a form of the well-studied online matching problem [23], current models developed in the EconCS, AI, and Operations Research communities [28,11,5] do not completely capture the mobile aspects of both the drivers and riders. Drivers and riders are agents who move about a constrained space (e.g., city streets), becoming (in)active periodically due to the matching process. When a platform receives a request, it must make a near-real-time dispatch decision amongst nearby drivers who are available at the current time. The platform's goal is to maximize an objective (e.g., revenue or throughput) by servicing requests in an online fashion, subject to various realworld constraints and challenges like setting prices, predicting supply and demand, fairness considerations, competing with other firms, and so on [8,17,4].In this paper, we study the dynamics of the nascent rideshare market under different dispatch strategies. Recent work uses Markov chains to model complex ride-sharing dynamics in a closed-world system-that is, a system with a fixed total supply of cars [28,4,5]. These assume the Markov chains reach their stationary distributions quickly, and thus all prior work optimizes for dispatch strategies in the limit. As a complement, the present paper characterizes-theoretically and empirically-when that limit assumption is valid under a variety of natural strategies.

show abstract

“…There are infinitely many λ * satisfying (7), but (9) shows that λ * i − λ * j is uniquely determined by l + ik , l + jk , and v k , k ∈ N . Next, we interpret l + ik and l + jk by introducing an electrical network.…”

Section: Value Of λmentioning

confidence: 99%

Analyzing Location-Based Advertising for Vehicle Service Providers Using Effective Resistances

Wei

Berry

2019

SIGMETRICS Perform. Eval. Rev.

View full text Add to dashboard Cite

Vehicle service providers can display commercial ads in their vehicles based on passengers' origins and destinations to create a new revenue stream. In this work, we study a vehicle service provider who can generate different ad revenues when displaying ads on different arcs (i.e., origin-destination pairs). The provider needs to ensure the vehicle flow balance at each location, which makes it challenging to analyze the provider's vehicle assignment and pricing decisions for different arcs. For example, the provider's price for its service on an arc depends on the ad revenues on other arcs as well as on the arc in question. To tackle the problem, we show that the traffic network corresponds to an electrical network. When the effective resistance between two locations is small, there are many paths between the two locations and the provider can easily route vehicles between them. We characterize the dependence of an arc's optimal price on any other arc's ad revenue using the effective resistances between these two arcs' origins and destinations. Furthermore, we study the provider's optimal selection of advertisers when it can only display ads for a limited number of advertisers. If each advertiser has one target arc for advertising, the provider should display ads for the advertiser whose target arc has a small effective resistance. We investigate the performance of our advertiser selection strategy based on a real-world dataset. 1 We use an example to estimate the value of the unit ad revenue. According to [1], the average cost per click (CPC) of Youtube ads was around $3.58 in the second quarter of 2018. Based on [14] and [24], the click-through rate (CTR) of an in-stream video ad was around 1.84% (note that this number was achieved without location-based targeting and the actual CTR of an in-vehicle video ad may be higher). When a provider displays 3 in-stream video ads to a user during each time slot (10 minutes), the unit ad revenue is $0.20 (= 3 × 1.84% × $3.58). This value is comparable to a vehicle service provider's net profit when it does not display ads. According to [17], Uber's net profit is around $8.23 per hour per vehicle, which corresponds to $1.37 per time slot per vehicle. 2 When a user takes the vehicle service, the corresponding consumer surplus is the difference between the highest price that the user is willing to pay for the service and the price that it actually pays.Location-Based Advertising 6:3 an associated electrical network. 3 Each location corresponds to a node in the electrical network. If there is a positive traffic demand between two locations, there is a resistor between the two corresponding nodes, and its resistance is determined by the traffic demand and travel time between the two locations. Given the associated electrical network, we can compute the effective resistance between any two nodes. Intuitively, a small effective resistance implies that there are many paths between the two locations and the provider can easily route vehicles between them. Based on the effective re...

show abstract

Empty-Car Routing in Ridesharing Systems

Cited by 181 publications

References 26 publications

A Probabilistic Approach for Demand-Aware Ride-Sharing Optimization

A Probabilistic Approach for Demand-Aware Ride-Sharing Optimization

Mix and Match: Markov Chains and Mixing Times for Matching in Rideshare

Analyzing Location-Based Advertising for Vehicle Service Providers Using Effective Resistances

Contact Info

Product

Resources

About