[Paper Review] Dynamic Pricing for Revenue Maximization in Mobile Social Data Market with Network Effects

Zehui Xiong, Dusit Niyato+3 more|arXiv (Cornell University)|Aug 13, 2018

Human Mobility and Location-Based Analysis34 references4 citations

TL;DR

This paper proposes a sequential dynamic pricing scheme for a mobile network operator to maximize long-term revenue in a social data market, leveraging network effects and accounting for wireless congestion. It demonstrates analytically and numerically that dynamic pricing outperforms static pricing in both operator revenue and user utility, with a fairness-aware modification ensuring equitable utility distribution among users.

ABSTRACT

Mobile data demand is increasing tremendously in wireless social networks, and thus an efficient pricing scheme for social-enabled services is urgently needed. Though static pricing is dominant in the actual data market, price intuitively ought to be dynamically changed to yield greater revenue. The critical question is how to design the optimal dynamic pricing scheme, with prospects for maximizing the expected long-term revenue. In this paper, we study the sequential dynamic pricing scheme of a monopoly mobile network operator in the social data market. In the market, the operator, i.e., the seller, individually offers each mobile user, i.e., the buyer, a certain price in multiple time periods dynamically and repeatedly. The proposed scheme exploits the network effects in the mobile users' behaviors that boost the social data demand. Furthermore, due to limited radio resource, the impact of wireless network congestion is taken into account in the pricing scheme. Thereafter, we propose a modified sequential pricing policy in order to ensure social fairness among mobile users in terms of their individual utilities. We analytically demonstrate that the proposed sequential dynamic pricing scheme can help the operator gain greater revenue and mobile users achieve higher total utilities than those of the baseline static pricing scheme. To gain more insights, we further study a simultaneous dynamic pricing scheme in which the operator determines the pricing strategy at the beginning of each time period. Mobile users decide on their individual data demand in each time period simultaneously, considering the network effects in the social domain and the congestion effects in the network domain. We construct the social graph using Erdős-Rényi (ER) model and the real dataset based social network for performance evaluation.

Motivation & Objective

• To address the gap in dynamic pricing models that jointly consider network effects and congestion in mobile social data markets.
• To design an optimal dynamic pricing policy that maximizes the long-term expected revenue for a monopoly mobile network operator.
• To ensure social fairness by modifying the pricing policy to balance individual user utilities.
• To compare the performance of sequential and simultaneous dynamic pricing schemes under realistic network conditions.

Proposed method

• The paper models the mobile social data market as a sequential decision process where the operator sets prices over multiple time periods based on user demand dynamics.
• It incorporates network effects via a social graph model (Erdős-Rényi and real-world datasets) to reflect how user demand increases with friends' activity.
• Congestion effects are modeled through a capacity-constrained radio access network, where increased user demand leads to higher delays and reduced utility.
• A modified sequential pricing policy is introduced to ensure fairness by adjusting prices based on individual user utility, preventing excessive burden on low-utility users.
• Theoretical analysis derives conditions under which the dynamic pricing scheme yields higher revenue and utility than static pricing.
• Numerical evaluation uses both synthetic (Erdős-Rényi) and real social network datasets to validate performance under varying network effect strengths and congestion levels.

Experimental results

Research questions

• RQ1 How does dynamic pricing compare to static pricing in terms of long-term revenue and user utility in a mobile social data market?
• RQ2 What is the impact of network effects on the operator’s revenue and user demand in dynamic pricing settings?
• RQ3 How do congestion effects in the wireless network domain constrain the effectiveness of dynamic pricing strategies?
• RQ4 Can a fairness-aware dynamic pricing policy ensure equitable user utility while maximizing operator revenue?
• RQ5 What are the performance differences between sequential and simultaneous dynamic pricing schemes in this context?

Key findings

• The proposed sequential dynamic pricing scheme achieves higher long-term revenue for the operator compared to static pricing, with numerical results showing significant improvement under strong network effects.
• Mobile users achieve higher total utility under dynamic pricing than under static pricing, due to better alignment of price with demand and network conditions.
• The dynamic pricing scheme effectively leverages network effects, where increased user activity leads to higher demand from connected users, amplifying overall market value.
• The fairness-aware modification of the pricing policy ensures that individual user utilities are balanced, preventing utility degradation for low-activity users.
• Performance evaluation on both Erdős-Rényi and real-world social network datasets confirms that dynamic pricing consistently outperforms static pricing across varying network topologies and congestion levels.
• The theoretical analysis proves that the dynamic pricing policy maintains non-negative revenue and utility gains under the given assumptions, with convergence to optimal pricing behavior.