Dynamic Pricing for Controlling Age of Information

Xuehe Wang, Lingjie Duan

Abstract

Fueled by the rapid development of communication networks and sensors in portable devices, today many mobile users are invited by content providers to sense and send back real-time useful information (e.g., traffic observations and sensor data) to keep the freshness of the providers' content updates. However, due to the sampling cost in sensing and transmission, an individual may not have the incentive to contribute the real-time information to help a content provider reduce the age of information (AoI). Accordingly, we propose dynamic pricing for the provider to offer age-dependent monetary returns and encourage users to sample information at different rates over time. This dynamic pricing design problem needs to balance the monetary payments to users and the AoI evolution over time, and is challenging to solve especially under the incomplete information about users' arrivals and their private sampling costs. For analysis tractability, we linearize the nonlinear AoI evolution in the constrained dynamic programming problem, by approximating the dynamic AoI reduction as a time-average term and solving the approximate dynamic pricing in closed-form. Then, we estimate this approximate term based on Brouwer's fixed-point theorem. Finally, we provide the steady-state analysis of the optimized approximate dynamic pricing scheme for an infinite time horizon, and show that the pricing scheme can be further simplified to an e-optimal version without recursive computing over time.

Introduction: The Challenge of Freshness

In the era of real-time applications like traffic navigation or environmental monitoring, the "freshness" of data is a critical performance metric. Researchers quantify this using Age of Information (AoI)—the time elapsed since the last data update was generated. Content providers typically rely on a "crowd" of mobile users to sense and transmit this data. However, users incur costs related to battery consumption and privacy, meaning they require incentives to participate.

System model of crowdsensing with dynamic pricing and transmission delay.

The provider faces a complex optimization problem characterized by incomplete information regarding user arrivals and private sampling costs. Furthermore, deciding the optimal price at every moment to maintain data freshness while minimizing payouts results in a "nonlinear dynamic program," which is notoriously difficult to solve due to the "curse of dimensionality."

Methodology: Achieving Tractability

The authors address these complexities through a three-step approximation framework:

  1. Linearization: The evolution of AoI is non-linear because expected age depends on the product of current age and the probability of a user accepting a price. The authors linearize this by replacing the complex dynamic term with a time-average estimator ($\delta$).
  2. Quadratic Value Functions: By assuming a quadratic cost for AoI—reflecting that very old data is significantly worse than slightly old data—they derive a closed-form solution for the optimal price $p(t)$.
  3. Fixed-Point Iteration: To ensure the time-average estimator ($\delta$) aligns with real-world outcomes, they employ Brouwer’s fixed-point theorem and an iterative algorithm to find a consistent value.
  4. Steady-State Analysis: Extending the model to an infinite time horizon allows for the identification of "steady-state" parameters. This leads to a simplified $\epsilon$-optimal pricing version that avoids recursive computing, making it practical for real-time server implementation.

Key Insights and Results

Numerical simulations over a 100-slot time horizon validate the theoretical model. Several key behaviors emerge from the optimized pricing strategy:

  • The Price of Freshness: The optimal price $p(t)$ is directly proportional to the current Age of Information $A(t)$. As data becomes "stale," the provider offers higher rewards to increase the likelihood of a user providing an update.
  • The "End of the World" Effect: In finite time horizons, the price offered to users drops toward zero as the deadline approaches. The provider avoids paying for fresh updates if the system is scheduled to shut down shortly thereafter.
  • Convergence and Optimality: Parameters $Q_t$ and $M_t$ converge rapidly to steady-state values. The performance gap between the complex optimal strategy and the simplified $\epsilon$-optimal strategy vanishes as the time horizon $T$ increases.
Evolution of optimal price, expected age, and convergence of parameters over time.

Conclusion and Future Directions

This work demonstrates that near-optimal pricing for AoI control can be achieved without excessive computational overhead. While the current model assumes a uniform distribution for user costs and single-user arrivals, future research could explore more complex scenarios. These include stochastic transmission delays in congested networks, multiple simultaneous user arrivals leading to auction-style mechanisms, and competitive environments where multiple providers vie for the same pool of mobile sensors.

Summary updated

Abstract

Fueled by the rapid development of communication networks and sensors in portable devices, today many mobile users are invited by content providers to sense and send back real-time useful information (e.g., traffic observations and sensor data) to keep the freshness of the providers' content updates. However, due to the sampling cost in sensing and transmission, an individual may not have the incentive to contribute the real-time information to help a content provider reduce the age of information (AoI). Accordingly, we propose dynamic pricing for the provider to offer age-dependent monetary returns and encourage users to sample information at different rates over time. This dynamic pricing design problem needs to balance the monetary payments to users and the AoI evolution over time, and is challenging to solve especially under the incomplete information about users' arrivals and their private sampling costs. For analysis tractability, we linearize the nonlinear AoI evolution in the constrained dynamic programming problem, by approximating the dynamic AoI reduction as a time-average term and solving the approximate dynamic pricing in closed-form. Then, we estimate this approximate term based on Brouwer's fixed-point theorem. Finally, we provide the steady-state analysis of the optimized approximate dynamic pricing scheme for an infinite time horizon, and show that the pricing scheme can be further simplified to an e-optimal version without recursive computing over time.
BibTeX
@inproceedings{wang2019dynamic,
  author = {Xuehe Wang and Lingjie Duan},
  title = {{Dynamic Pricing for Controlling Age of Information}},
  booktitle = {ISIT},
  year = {2019},
  url = {https://openreview.net/forum?id=SjKRgVGoZc},
}
Figure