关键词: 动态定价, 多摇臂赌博机, UCB1算法

Abstract:

To solve the single product pricing problem of online retailers with incomplete demand information, a product pricing algorithm based on multi-armed bandit is proposed. In order to improve the effect of multi-armed bandit algorithm in the pricing problem, the algorithm takes advantage of the monotonicity of demand curve and adds consumer preference recognition. This paper analyzes the reserve price of consumers to obtain the purchase probability of consumers, models the pricing problem of online retailers into a multi-armed bandit model, presents the corresponding pricing algorithm and makes a theoretical analysis, and finally compares the pricing effect of relevant algorithms through simulation experiments. Simulation results show that the algorithm improves the revenue of online retailers.

Key words: dynamic pricing, multi-armed bandit, UCB1 algorithm