Nov 26, 2020|

JDD Series: JD’s End-to-End Replenishment System Endorsed by Top Industry Journal


by Ling Cao

During JD’s fourth annual tech summit JD Discovery (JDD) held on November 25th in Beijing, JD has shared plans to build a digital and intelligent supply chain. JD-Y, JD’s supply chain R&D unit (focused on supply chain innovation) shared a relevant breakthrough and application regarding its self-built industry-leading end-to-end replenishment model (E2E model).

A paper on the model has been accepted and to be published soon by Management Science, a world-leading scholarly journal within INFORMS, an international association for professionals in operations research and analytics, which publishes scientific research on the practice of management. The journal has also been named top 20 journals valued by business school deans and academic program directors from BusinessWeek.

Dr. Max Shen, supply chain chief scientist of shared, “Our E2E model shortens the decision process and provides an automatic inventory management solution with the potential to generalize and scale. The concept of E2E, which uses the input information directly for the ultimate goal, can also be useful in practice for other supply chain management.”

Dr. Max Shen, supply chain chief scientist of

Dr. Max Shen

Specifically, the model can realize self-learning. Applying deep learning technology in JD’s real scenarios enables the company to achieve automatic merchandise replenishment based on historical sales performance without the need for sales forecast.

JD tested this model with the oil and condiments categories. After monitoring the effects for four months, JD saw that online product availability improved 0.5%, while the inventory turnover days reduced 2.8.

The future of supply chain is demand driven, open, synergetic, and agile. JD aims to build up and open its capabilities with industry partners. This model is a best practice example of how frontier technology can be applied in real scenarios.


To get the full paper, pls contact: