This project focused on the core technical problem of the multi‐factory production planning problem, which can be regarded as a relaxed linear programming problem of the large‐scale mixed‐integer programming problem. However, the existing solvers can hardly meet the time efficiency requirements, which urgently leads us to design a new numerical software.
After testing a brunch of popular algorithms for solving large-scale linear programming problems, we find that the simplex method significantly outperforms others on the real data sets, especially when we can obtain a good starting point. Motivated by these findings, we mainly focus on obtaining a good starting point and designing an efficient presolving strategy in this collaborative project. Based on the specific structure and data characteristics of the specific problems, our research team designed an efficient and robust algorithm for providing effective initial solutions which can contribute to efficient presolving methods. Consequently, combining the designed algorithm with the simplex methods, our team designed a software package that outperforms the current commercial software package for solving LP arising from the multi-factory production planning problem. The research outcome has greatly enhanced the efficiency of supply chain management and planning in daily production.
In recognition of the contributions made by our research team, Professor Sun Defeng has been awarded the Distinguished Collaborator Award both from Hong Kong Research Center and Huawei Noah’s Ark Lab, Huawei Technologies Co. Ltd. However, there are still many detailed issues that we need to solve. The efficiency and stability of the software package still have room for further improvement. Nowadays, the main focus of our research team is to design highly efficient solvers to solve practical application problems.
