Project description

Aiming at the problem of service level characterization, from the perspectives of inventory level and service fill rate, establish a service level index based on risk measure. Use distributionally robust optimization theory to model uncertain factors, and combine multi-product features and properties to design efficient algorithms.

Impact, Challenges

Impact: Enrich the theoretical research of IRP and promote the development of the VMI model.

Challenges: Characterize the satisficing measure of inventory level and demand fill rate; Construct the distributionally uncertainty set of random variables; Design an efficient algorithm based on the distributionally robust optimization.

Solution description

Research on Satisficing Measure Based on the Distributionally Robust Optimization.

Solution description

Inventory level target-oriented IRP

Demand fill rate target-oriented IRP

Algorithm Design and Analysis of IRP

Key contribution

Combine the characteristics of IRP to construct tractable measures to describe service level goal; 

Without the support of a large amount of data, develop an IRP model that can portray uncertain factors without increasing computational complexity;

Based on the characteristics of the model, design an efficient algorithm to solve the model.

Next steps

Apply our model to industry. 

Collaborators

Professor Li Xiaobo， NUS.

Team

Lianmin Zhang、Research Group in Shenzhen Research Institute of Big Data.