MUMBAI, India, Aug. 12 -- Intellectual Property India has published a patent application (202641094414 A) filed by Prof. Brijesh Singh; Cmr Institute Of Technology; Mr. Yashvir Singh; Mrs. P. R. Madhu Shree; Dr. Rupali Khaire; Mathu K; Rekha Gupta; and Dr. Kartik Chandra Das on August 04, 2026, for Reinforcement Learning For Dynamic Pricing Strategy In E-Commerce Platforms.
Inventors include Prof. Brijesh Singh; Mr. Yashvir Singh; Mrs. P. R. Madhu Shree; Dr. Rupali Khaire; Mathu K; Rekha Gupta; and Dr. Kartik Chandra Das.
The application for the patent was published on August 07, 2026, under issue no. 32/2026.
Abstract: ABSTRACT REINFORCEMENT LEARNING FOR DYNAMIC PRICING STRATEGY IN E-COMMERCE PLATFORMS This invention presents a reinforcement learning-based system for dynamic pricing in e-commerce platforms. The pricing problem is modeled as a Markov Decision Process, where the agent interacts with a simulated environment to learn optimal pricing policies. The state space incorporates product inventory, competitor prices, customer demand signals, time, seasonality, and price elasticity, while the action space consists of discrete price adjustments. The reward function is designed to maximize cumulative revenue while penalizing stockouts, margin erosion, and customer churn. Deep Q-Network (DQN) and Proximal Policy Optimization (PPO) algorithms are implemented and trained using historical transaction and clickstream data from 2023–2025. Results demonstrate that PPO achieved superior adaptability in volatile categories such as electronics and fashion, while DQN performed effectively in stable categories. The system integrates fairness constraints to prevent excessive price discrimination, ensuring regulatory compliance. This invention provides a robust, autonomous, and ethically aligned solution for real-time dynamic pricing optimization in e-commerce.
Disclaimer: Curated by HT Syndication.
