Research map / Learning theory and algorithms / Fairness
FairPlay: A Multi-Sided Fair Dynamic Pricing Policy for Hotels
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's full text. Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Proposes FairPlay, a dynamic hotel room pricing policy that uses Owen values from a cooperative game (graph-based induced subgraph game) to set prices and hotel exposure fairly for customers and providers, evaluated by simulation on Athens, Barcelona and Rome hotel data.
Problem
Overtourism and overpricing by hotels make accommodation unaffordable; existing dynamic pricing policies rarely consider fairness for customers and providers on two-sided platforms.
Approach
Model room pricing as a Dynamic Hotel-Rooms Game (a cooperative game with per-hotel coalitions), represent it as an induced subgraph game so Owen values are tractable, set prices as a minimum expected price times (1 + Owen value), and expose hotels via an Exposure-Owen ratio.
Key results
- FairPlay yields dynamic margins (e.g., 51.5% on Day 5 vs. static 17%) and the static policy overprices triple rooms on 20 of 30 days in the Athens large dataset.
- Hotel exposure rankings correlate with hotels' collective room power at 86.88% to 93.34% similarity across cities and dataset sizes (Table 2).
- Higher-quality hotels (higher booking probability) receive larger profit margins, while lower-quality hotels get tighter margins.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| Use the Exposure-Owen ratio to allocate hotel exposure opportunities | FairPlay · Athens, Barcelona, and Rome small, medium, and large datasets | ranking similarity: 90.48%; 93.34%; 86.88%; 89.71%; 89.17%; 87.5%; 88.04%; 89.89%; 89.18% | Helped |
| Replace demand-responsive pricing with a static 17% MEP margin | FairPlay · Hotel room pricing simulations | pricing fairness | Hurt |
| Set room prices using demand-responsive FairPlay instead of a static margin | Hotel room pricing · Athens large, hotel AT 0004, triple room, Day 5 | profit margin: 51.5% vs 17% | Mixed |
| Increase the probability that customers book at the higher-quality hotel | FairPlay · Two-hotel simulated datasets | provider profit margin | Helped |
Surprising findings
- A sole remaining hotel with available rooms of a type, when all others are sold out, gets zero profit margin under FairPlay.
Limitations
- Evaluation is purely simulation-based with simulated reservation and cancellation processes rather than real booking behavior.
- Only room type is used as a dependency feature; other features such as stars, ratings, and breakfast are left to future work.
- Static 17% baseline value was chosen empirically, and the fairness claims rely on the authors' own definition of multi-sided fairness.
Experimental setup
Datasets: Athens hotel room data (small/medium/large) Barcelona hotel room data Rome hotel room data synthetic two-hotel datasets (new small and large)
Metrics: price vs. minimum expected price profit margin rank similarity (Kendall tau style) between exposure and col profit margins per hotel under varying quality
Baselines: Static 17% margin pricing policy
Builds on: Owen value Shapley value induced subgraph games FairRec Equity of attention
Closest papers by meaning
- Dynamic Service Fee Pricing under Strategic Behavior: Actions as Instruments and Phase Transition (NeurIPS 2024 · similarity 0.81)
- Homogeneous Algorithms Can Reduce Competition in Personalized Pricing (NeurIPS 2025 · similarity 0.81)
- Interpolating Item and User Fairness in Multi-Sided Recommendations (NeurIPS 2024 · similarity 0.80)
- Decongestion by Representation: Learning to Improve Economic Welfare in Marketplaces (ICLR 2024 · similarity 0.79)
- Fair Allocation in Dynamic Mechanism Design (NeurIPS 2024 · similarity 0.79)
- Mixed Fair Division: A Survey (AAAI 2024 · similarity 0.79)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.