Optimizing Digital Coupon Assignment Using Constrained Reinforcement Learning

Abstract

Coupon marketing is a traditional but effective way to retain customers and stimulate new purchases. Recently, digital coupons have been widely used in e-commerce and distributed to almost everyone. However, the decision on when and whom to issue the coupon is often based on managers' experience and calling for optimization and automation. Collaborated with a leading e-commerce platform, we propose an exploratory constrained reinforcement learning modeling to optimize digital coupon distribution policy under the constraint of maximum offering number. Our experimental results showed that the optimal policy could increase the cumulative total sales about 6% comparing to the original policy of the platform. This work enriches the applications of reinforcement learning in real-world business practices and provides useful implications for future study on constrained reinforcement learning.