Joint optimization strategy for dynamic pricing and customer behavior in cross-border E-commerce based on deep reinforcement learning
Abstract
To address the insufficient adaptability between dynamic pricing and customer behavior in cross-border e-commerce, this paper proposes a joint optimization strategy based on deep reinforcement learning (DRL). A model centered on the Proximal Policy Optimization (PPO) algorithm is constructed, integrating multi-dimensional variables such as commodity costs, exchange rates, tariffs, and customer price sensitivity. It designs state spaces, action spaces, and weighted reward functions suitable for cross-border scenarios, and completes model training and optimization through data preprocessing and hyperparameter tuning. Empirical comparisons with logistic regression-based static pricing models and traditional DRL pricing models (DQN) show that the proposed model increases profit margin by 15.3%, improves inventory turnover efficiency by 27.9%, and achieves better pricing stability. This study verifies the effectiveness of DRL in joint optimization, providing technical support and practical references for cross-border e-commerce enterprises to implement precise pricing and enhance customer retention.