Yingming Mao

In Huawei Company, I was engaged in the optimization of the memory waterline of the HarmonyOS operating system and accumulated a lot of work experience.

In Meituan Company, I was responsible for the vehicle transportation planning issues in the Youxuan Business Division. By utilizing the fast algorithm designed by me, Meituan Company could obtain a vehicle routing plan with a fewer number of required vehicles within an extremely short period of time.

• Piecewise Linearization Design: Developed an optimal piecewise strategy and corresponding linearization coefficients using mixed-integer programming within a complete orthogonal decomposition least squares framework, incorporating SVM principles.
• Algorithm Testing: Compared with methods in IEEE Transactions on Power Systems, achieving a 29.8% reduction in average error and a 22.2% reduction in maximum error on IEEE standard cases.

• Data Generation Method: Established a method involving parameter estimation, sampling, baseline setting, and data computation. Employed multivariate normal distribution assumptions, eigenvalue decomposition, and Cholesky decomposition to accurately generate data.
• Algorithm Testing: The proposed sampling method significantly reduced the data-driven model error by over 90%, with the baseline setting method achieving similar improvements. This work has been published in Energy Reports, SCI Zone 2, Impact Factor 4.937.

• Model Construction: Built an optimization model based on graph theory considering network topology, nodes, hop count, latency constraints, and business demands.
• Offline-Online Framework: In the offline framework, introduced a scheduling method based on Lagrangian relaxation, achieving a 98% scheduling rate and a 15% improvement in load balancing efficiency over traditional methods. Implemented rapid matching calculations in the online framework, achieving second-level solutions for 100 nodes and 1000 services.