Abstract: In recent years, more and more attention has been paid to carbon emission and sustainable development of supply chain. Returnable transport items(RTIs) have been widely concerned in academia and enterprises. However, the number of RTIs after recycling is less than the quantity when it is sent out, which not only causes the reduction of enterprise assets, but also may cause the shortage of downstream manufacturers. At present, different methods have been considered to reduce the number of RTIs lost, but no study has considered the effect of learning effect on the reduction of RTIs loss rate. Therefore, based on the consideration that recycling of RTIs is manually operated and the work content is highly repetitive, it is assumed that the learning effect of workers can reduce the discarding rate of RTIs, In order to minimize the total cost of supply chain, two models are constructed. The first model does not consider the learning effect, the discarding rate of reusable containers is a random variable, and the second model considers the learning effect, and the discarding rate of reusable containers decreases continuously. By comparing the two models, the influence of learning effect on RTIs discarding rate and total cost of supply chain is discussed.

Key words: closed-loop supply chain; returnable transport items; stochastic RTI return quantities; learning effect