Joint Pricing and Matching in Mobility Networks

39 concepts24 papers1 proposal

In traditional on-demand transportation platforms, two fundamental decisions are typically made in isolation: pricing (the economic lever used to manage demand) and matching (the operational lever used to dispatch vehicles and route supply).

When a platform experiences a surge in ridership, the pricing algorithm raises fares to deter excess demand and attract drivers. Moments later, a separate dispatching algorithm pairs the remaining passengers with available vehicles based purely on proximity or travel time.

While computationally convenient, this decoupled approach creates a fundamental mismatch. Price signals directly dictate the volume and spatial distribution of incoming requests, while matching decisions determine the physical availability and operational costs of the fleet. Because these two forces are deeply interdependent, separating them prevents platforms from achieving global efficiency, often leading to long passenger wait times, strategic driver rejections, and empty vehicle miles.

To solve this, a growing body of research focuses on the unified optimization of supply and demand levers. By treating pricing, matching, routing, and fleet rebalancing as a single integrated problem, platforms can synchronize economic incentives with physical capacity constraints.


The Failure of Decoupled Systems

Why does separating pricing from matching cause sub-optimal performance? The core issue lies in a cyclic dependency: you cannot determine the optimal price for a trip without knowing the operational cost of fulfilling it, and you cannot efficiently allocate a vehicle without knowing how pricing will alter future demand.

As argued by Nikhil Garg et al. in Driver Surge Pricing, treating dispatching and fare setting as independent problems creates misaligned incentives. For instance, multiplicative surge pricing can fail to account for the true opportunity cost of a driver's time, prompting drivers to strategically reject short trips.

Furthermore, independent pricing mechanisms fail to recognize how fares influence physical vehicle distribution. Mengjing Chen et al. demonstrate in Dispatching Through Pricing: Modeling Ride-Sharing and Designing Dynamic Prices that price signals act as implicit routing mechanisms for drivers. When platforms decouple these choices, they fail to capture how price adjustments alter passenger acceptance and driver trajectories, leading to severe spatial imbalances where vehicles end up stranded far from active demand.

This friction is not unique to ride-hailing. Even in scheduled transit systems, such as suburban rail networks, optimizing operations and fare structures independently causes systemic inefficiencies. As shown by Pengling Wang et al. in Fare system selection and line planning for new suburban rail lines considering the supply–demand interactions, pricing decisions directly dictate ridership volume, which in turn necessitates specific operational capacities and line plans.


The Cyclic Dependency of Mobility Platforms

To build an efficient transport network, the feedback loop between demand generation and supply allocation must be resolved simultaneously.

┌────────────────────────────────────────┐
│          Pricing Decisions             │
└───────────────────┬────────────────────┘
                    │ Influences
                    ▼
┌────────────────────────────────────────┐
│       Passenger Demand Volume          │
└───────────────────┬────────────────────┘
                    │ Constrains
                    ▼
┌────────────────────────────────────────┐
│      Matching & Routing Capacity       │
└───────────────────┬────────────────────┘
                    │ Determines
                    ▼
┌────────────────────────────────────────┐
│      Future Vehicle Availability       │
└───────────────────┬────────────────────┘
                    │ Feeds Back Into
                    ▼
                    [Loop Back to Pricing]

As formalized by Pradeep Varakantham et al. in Future Aware Pricing and Matching for Sustainable On-Demand Ride Pooling, this cyclic dependency means that isolated matching algorithms often allocate resources to low-value or highly inefficient requests, draining the fleet's future capacity.

Kshipra Bhawalkar et al. reinforce this principle in Revenue Maximization in Transportation Networks, proving that an effective network strategy must account for how price changes alter customer demand patterns, and how immediate vehicle availability dictates the feasibility of serving those trips at the proposed prices.


Pricing as a Demand-Throttling Filter

The core intuition behind joint optimization is that pricing should serve as a meta-optimization layer for matching. Instead of treating pricing merely as a revenue generator, integrated frameworks use price as a filter to shape demand so that it perfectly matches the combinatorial constraints of the physical fleet.

In Joint Pricing and Matching for City-Scale Ride-Pooling, Sanket Shah et al. leverage auction theory to design a posted-price mechanism that naturally throttles demand during peak periods. By aligning customer acceptance with actual vehicle capacity, the system prevents the platform from over-promising rides that the fleet cannot fulfill, maintaining high service reliability.

Mathematically, this integration transforms how platforms calculate expected revenue. Yuya Hikima et al. demonstrate in Integrated Optimization of Bipartite Matching and Its Stochastic Behavior: New Formulation and Approximation Algorithm via Min-cost Flow Optimization that simultaneously optimizing control incentives and bipartite matching decisions yields significantly higher expected revenue than multi-stage, independent processing. By mapping the joint problem to a convex min-cost flow optimization, they provide provable approximation guarantees for real-time dispatching environments.


Unifying Physical Fleet Logistics

When supply and demand management are unified, the platform can seamlessly trade off the economic costs of demand manipulation against the physical costs of fleet logistics, such as vehicle rebalancing and repositioning.

Rebalancing as a Service Request

In autonomous mobility-on-demand (AMoD) systems, empty vehicles must be constantly relocated to anticipate future demand. Ioannis Paschalidis et al. show in Joint Pricing and Rebalancing of Autonomous Mobility-on-Demand Systems and Optimal Operations Management of Mobility-on-Demand Systems that formulating pricing and vehicle relocation as a single unified quadratic program allows the system to dynamically balance the costs of moving empty cars against the revenue impacts of price adjustments, boosting platform profitability by up to 50%.

To execute this in real time, Zeno Woywood et al. propose in Multi-agent Soft Actor-Critic with Coordinated Loss for Autonomous Mobility-on-Demand Fleet Control a framework that treats proactive rebalancing actions as a special category of customer requests. This allows a single matching mechanism to handle both passenger fulfillment and vehicle repositioning simultaneously.

Spatio-Temporal State Transitions

By framing both active order dispatching and idle driver repositioning as state transitions within a shared objective function, the platform ensures that immediate passenger fulfillment does not compromise long-term fleet distribution.


Expanding the Optimization Horizon

As mobility platforms grow more complex, joint optimization frameworks are expanding to handle multi-modal operations and strict temporal constraints.

For instance, the PassGoodPool framework, introduced by Kaushik Manchella et al. in PassGoodPool: Joint Passengers and Goods Fleet Management With Reinforcement Learning Aided Pricing, Matching, and Route Planning, applies joint reinforcement learning to fleets that transport both passengers and goods. By co-optimizing pricing, matching, and multi-hop route planning, the system maximizes vehicle utilization across completely different demand profiles.

Similarly, joint optimization can resolve the reliability requirements of daily commuters. In The Commute Trip-Sharing Problem, Mohd. Hafiz Hasan et al. model transportation requests as coupled pairs of inbound and outbound journeys. Treating the daily commute as a single balanced problem rather than independent legs ensures a guaranteed return trip for participants, reducing overall vehicle usage by up to 57% while maintaining strict service guarantees.

Ultimately, whether managing autonomous fleets, mixed-delivery networks, or urban rail systems, the shift away from decoupled, siloed algorithms toward joint supply-demand optimization allows platforms to replace reactive corrections with proactive, globally coordinated efficiency.

Key Foundational Work

A Profit-Aware Negotiation Mechanism for On-Demand Transport Services (2014) — contributed 2 concept elements to this direction.
Dispatching Through Pricing: Modeling Ride-Sharing and Designing Dynamic Prices (2019) — contributed 1 concept elements to this direction.
Deep Reinforcement Learning for Ride-sharing Dispatching and Repositioning (2019) — contributed 1 concept elements to this direction.
History & citationVersion 1 · 2026-10-04

Current · Permanent link

View frozen version 1

Cite this direction

Latest version (a living page):

@misc{lacuna_joint_supply_demand_optimization_07c3ec3f,
  title = {joint supply-demand optimization},
  author = {{Lacuna}},
  howpublished = {Lacuna research direction},
  year = {2026},
  url = {https://lacuna.tiptreesystems.com/direction/joint-supply-demand-optimization/rdr_07c3ec3f179047d4b2338438ab81b937},
  note = {Latest version of a living research direction; cite a version for a fixed text. Accessed 2026-11-19},
}

This exact version (version 1, frozen text):

@misc{lacuna_joint_supply_demand_optimization_07c3ec3f_v1,
  title = {joint supply-demand optimization},
  author = {{Lacuna}},
  howpublished = {Lacuna research direction},
  year = {2026},
  url = {https://lacuna.tiptreesystems.com/direction/joint-supply-demand-optimization/rdr_07c3ec3f179047d4b2338438ab81b937/v/1},
  note = {Version 1, frozen 2026-10-04. Accessed 2026-11-19},
}
About

These directions overlap with this one through shared papers.

See something wrong? Report a mistake