How gamzetuncay/dynamicpricingwithreinforcementlearning works
This repository is a research implementation of reinforcement-learning-based dynamic hotel pricing. The documented workflow models pricing as an MDP, uses Q-learning and a reward balancing profit with demand, and uses digitized hotel features to support estimates for newly added hotels. The notebooks indicate preprocessing, model experiments, and result analysis; the README does not map those stages to specific notebooks, and no source excerpts were provided, so edges show only documented relationships.
Main entry points: Dynamic_Pricing_3acts_modular.ipynb Pricing experiments · Q_table.ipynb Q-learning · Preprocessing.ipynb Data preprocessing · Dynamic_Pricing_3acts_modular_train - validation100.ipynb Training validation
Components 14
Pricing model
- Pricing MDP
Dynamic_Pricing_3acts_modular.ipynb - Profit-demand reward
Dynamic_Pricing_3acts_modular.ipynb - Q-learning
Q_table.ipynb
Hotel similarity
- Hotel features
Dynamic_Pricing_3acts_modular.ipynb - Hotel similarity
Dynamic_Pricing_3acts_modular.ipynb - New hotel estimates
Dynamic_Pricing_3acts_modular.ipynb
Experiments and analysis
- Data preprocessing
Preprocessing.ipynb - Pricing experiments
Dynamic_Pricing_3acts_modular.ipynb - Training validation
Dynamic_Pricing_3acts_modular_train - validation100.ipynb - Hotel experiments
Dynamic_Pricing_3acts_modular_noncurriculum_byhotels.ipynb - Week experiments
Dynamic_Pricing_3acts_modular_noncurriculum_byweeks.ipynb - Results analysis
Figures.ipynb
Other components
- Researcher
- Hotel sales data
Connections 11
- Researcher → Pricing experiments: runs experiments
README.md - Hotel sales data → Data preprocessing: provides records
README.md - Hotel features → Hotel similarity: supports comparisons
README.md - Hotel similarity → New hotel estimates: enables estimates
README.md - Pricing MDP → Q-learning: solved with
README.md - Profit-demand reward → Q-learning: guides learning
README.md - Hotel sales data → Pricing experiments: evaluates predictions
README.md - Pricing experiments → Training validation: evaluated by
- Pricing experiments → Hotel experiments: varied by hotel
- Pricing experiments → Week experiments: varied by week
- Pricing experiments → Results analysis: results analyzed