Data Analyst Intern

Internship, Lenovo, Solution and Service Group, 2022-07-11 - 2022-10-10

From Jul. 2022 to Oct. 2022, I joined the Solution & Services Group at Lenovo as a data analyst intern. I participated in the development of several systems, performing Python development, code review, system testing, and document maintenance.


MARS

MARS Core Solver: multi-period, multi-priority, multi-objective

MARS is a forecasting system for chipset inventory. Using the data of end-product demand, component supply, and bill-of-materials, the system optimizes for the maximum production output and minimum latency while considering the priority and balance in production volume between different end-products. The system employs the simplex method for solving the optimum using the GLPK solver in the Cvxpy module. It outputs the shortages of components in different periods.

Our team works on the solver/algorithm of MARS. During its development, I reviewed the solver’s code, maintained the Confluence document, and conducted testing of MARS, during which I fixed bugs in LP constraints, streamlined the constraint matrix of the linear programming model, and reduced solution time by 40%.

Surface Mount Technology Optimization System

SMT Optimization system is a production planning system for PCBA assembly, which optimizes for minimum PCB waste and optimum priority sequence.

During the prototyping phase, I developed a preliminary solution for the system, constructing multi-objective mixed-integer programming models using the GLPK solver in the Cvxpy module to optimize the number of PCB panels produced. I independently coded the demo and presented the solution.

Smart Advanced Planning System

Smart APS system is a production scheduling system for planning the order sequences on different production lines with each order having different priority and production constraints.

I developed the demo for the system, modeling the problem into a multi-objective mixed-integer programming model, similar to the vehicle routing problem and parallel machine scheduling problem. The system implemented the genetic algorithm using the Geatpy module.

Knowledge Share on Python

Maintained documents on the Confluence platform. Wrote knowledge-share documents and hosted a knowledge-share event about Python.