Abstract:The use of dynamic pricing by profit-maximizing firms gives rise to demand fairness concerns, measured by discrepancies in consumer groups' demand responses to a given pricing strategy. Notably, dynamic pricing may result in buyer distributions unreflective of those of the underlying population, which can be problematic in markets where fair representation is socially desirable. To address this, policy makers might leverage tools such as taxation and subsidy to adapt policy mechanisms dependent upon their social objective. In this paper, we explore the potential for AI methods to assist such intervention strategies. To this end, we design a basic simulated economy, wherein we introduce a dynamic social planner (SP) to generate corporate taxation schedules geared to incentivizing firms towards adopting fair pricing behaviours, and to use the collected tax budget to subsidize consumption among underrepresented groups. To cover a range of possible policy scenarios, we formulate our social planner's learning problem as a multi-armed bandit, a contextual bandit and finally as a full reinforcement learning (RL) problem, evaluating welfare outcomes from each case. To alleviate the difficulty in retaining meaningful tax rates that apply to less frequently occurring brackets, we introduce FairReplayBuffer, which ensures that our RL agent samples experiences uniformly across a discretized fairness space. We find that, upon deploying a learned tax and redistribution policy, social welfare improves on that of the fairness-agnostic baseline, and approaches that of the analytically optimal fairness-aware baseline for the multi-armed and contextual bandit settings, and surpassing it by 13.19% in the full RL setting.

What problem does this paper attempt to address?

### What problems does this paper attempt to solve? This paper aims to solve the fairness problems caused by dynamic pricing in the market. Specifically, dynamic pricing may lead to differences in demand responses among consumer groups, resulting in under - representation of certain groups in the market and thus unfairness. For example, in markets such as health insurance and housing, dynamic pricing may lead to lower participation of specific ethnic or socioeconomic groups, further exacerbating existing social inequalities. To meet this challenge, the author proposes the following research objectives: 1. **Introduce a new policy mechanism framework**: Design three different policy mechanisms (multi - armed bandit, contextual multi - armed bandit, and reinforcement learning) to solve the demand fairness problem in a single - product market. 2. **Implement multiple fairness incentive mechanisms**: Use a series of economic policy variants (such as taxes and subsidies) to encourage enterprises to consider fairness when pricing. 3. **Propose FairReplayBuffer**: Design a special replay buffer for the Soft Actor - Critic (SAC) algorithm to ensure that RL agents can evenly sample experiences from different fairness intervals, so as to better handle rare but important events. 4. **Combine subsidies and taxes**: Propose a comprehensive method to effectively solve the demand fairness problem and improve social welfare by combining taxes and subsidies. 5. **Comprehensively evaluate the proposed framework**: Evaluate the effect of the proposed framework through simulation studies, involving the performance of multiple enterprises in the face of two different consumer behaviors. ### Research background and problem description Dynamic pricing is a personalized pricing strategy based on consumers' willingness to pay. Although it can improve the profitability and sales speed of enterprises, it may also bring negative social impacts, especially in markets such as insurance and housing, which may lead to a decline in the participation of specific groups and further exacerbate social inequalities. For example, Hispanic people in the United States are more difficult to obtain medical care coverage than black, white, and Asian people, which may be an unfair result caused by dynamic pricing strategies. In addition, dynamic pricing may also lead to consumers' negative perception of fairness, affecting market prospects and buyer participation, thereby spreading existing gaps. Therefore, this article explores how to improve this unbalanced buyer distribution problem by introducing fairness incentive mechanisms, especially in those markets where the buyer distribution should reflect the basic population characteristics. ### Solution overview To solve the above problems, this paper proposes a framework led by a dynamic Social Planner (SP), which uses AI methods to generate tax and subsidy policies to encourage enterprises to adopt more fair pricing behaviors. Specifically: - **Non - linear programming (NLP)**: Used to show how enterprises adjust price distribution when considering fairness, usually only sacrificing a small amount of profit. - **Dynamic Social Planner (SP)**: Generate tax plans to punish unfair enterprise behaviors and redistribute tax revenues to under - represented consumer groups to encourage higher market participation. - **Reinforcement learning (RL)**: Apply the Soft Actor - Critic (SAC) algorithm to train agents so that they can optimize tax and subsidy strategies in different situations and maximize social welfare. Finally, the study found that by deploying the learned tax and redistribution policies, social welfare has been significantly improved, approaching or even exceeding the theoretically optimal fairness baseline. ### Summary This paper explores how to promote market fairness through policy tools (such as taxes and subsidies) in the context of dynamic pricing by introducing AI and reinforcement learning technologies, thereby improving overall social welfare.

Fairness Incentives in Response to Unfair Dynamic Pricing

Spatial-temporal Pricing for Ride-Sourcing Platform with Reinforcement Learning

The AI Economist: Improving Equality and Productivity with AI-Driven Tax Policies

Utility Fairness in Contextual Dynamic Pricing with Demand Learning

Fairness-aware Online Price Discrimination with Nonparametric Demand Models

Fairness, Welfare, and Equity in Personalized Pricing

How is the AI Perceived when It Behaves (Un)fairly?

The AI Economist: Taxation policy design via two-level deep multiagent reinforcement learning

The AI Economist: Optimal Economic Policy Design via Two-level Deep Reinforcement Learning

Long-term Fairness in Ride-Hailing Platform

TaxAI: A Dynamic Economic Simulator and Benchmark for Multi-Agent Reinforcement Learning

Dynamic pricing strategy for content products considering consumer fairness concerns and strategic behavior

Fairness-Aware Dynamic Ride-Hailing Matching Based on Reinforcement Learning

Grace Period is All You Need: Individual Fairness without Revenue Loss in Revenue Management

AI-driven Prices for Externalities and Sustainability in Production Markets

Designing Fair Systems for Consumers to Exploit Personalized Pricing

Fair Incentives for Repeated Engagement

Fair Dynamic Rationing

Implementing Fairness Constraints in Markets Using Taxes and Subsidies

Behavior-Based Pricing: An Analysis of the Impact of Peer-Induced Fairness

Building a Foundation for Data-Driven, Interpretable, and Robust Policy Design using the AI Economist