Dynamic Pricing

Developers of our company have implemented an algorithm for prediction of optimal prices for online store products. The prediction algorithm was based on earlier sales data of similar goods, cost of goods and data about competitor sales for the given period.

Using the method of reinforced learning, the Q-learning program selects an optimal strategy for price changes (increase or decrease) depending on the response of users in order to increase store profits. According to customer’s requirements our developers have chosen necessary input data, system settings and optimization metrics in order to create and test the model.

The algorithm supports an ability to select products from a database in order to predict the possibility of retraining the system for new data and an ability to configure the retraining process automatically.