Operating Range Estimation for Price Optimization

We focus on price range estimation in price optimization. In recent years, price optimization methods have become crucial for maximizing revenues and profits; however, it is still challenging to apply them to real pricing operations. Operators should take responsibility for setting prices, and it is essential for the price optimization methods to offer a reasonable price range that can be easily interpreted and accepted by them. However, most papers have considered price ranges as given constant values, and only a few papers have discussed reasonable price ranges in real applications. In this paper, we propose a price range estimation method using association rules and mathematical optimization. Specifically, we estimate the price range for each product from the previously set price using association rules with historical pricing data. We then adjust the price range by solving an optimization problem with shape constraints. The proposed method achieves a price optimization under a reasonable price range for the operators in real applications. We conduct numerical experiments using historical pricing data in real services and demonstrate the effectiveness of the proposed method.