Online Pricing Incentive to Sample Fresh Information

TL;DR

This paper introduces an online pricing strategy to incentivize mobile drivers to sample diverse paths, thereby controlling the age of information (AoI) in a dynamic, uncertain environment.

Contribution

It develops the first online pricing framework for incentivizing path diversity in AoI management, including algorithms for single and multiple path scenarios.

Findings

Optimal pricing algorithm for single non-shortest path based on MDPs.

Path sampling order affects AoI control effectiveness.

Approximation algorithm performs closer to optimal with more paths.

Abstract

Today mobile users such as drivers are invited by content providers (e.g., Tripadvisor) to sample fresh information of diverse paths to control the age of information (AoI). However, selfish drivers prefer to travel through the shortest path instead of the others with extra costs in time and gas. To motivate drivers to route and sample diverse paths, this paper is the first to propose online pricing for a provider to economically reward drivers for diverse routing and control the actual AoI dynamics over time and spatial path domains. This online pricing optimization problem should be solved without knowing drivers' costs and even arrivals, and is intractable due to the curse of dimensionality in both time and space. If there is only one non-shortest path, we leverage the Markov decision process (MDP) techniques to analyze the problem. Accordingly, we design a linear-time algorithm for returning optimal online pricing, where a higher pricing reward is needed for a larger AoI. If there are a number of non-shortest paths, we prove that pricing one path at a time is optimal, yet it is not optimal to choose the path with the largest current AoI. Then we propose a new backward-clustered computation method and develop an approximation algorithm to alternate different paths to price over time. Perhaps surprisingly, our analysis of approximation ratio suggests that our algorithm's performance approaches closer to optimum given more paths.