Case Study: VGEN achieves 50% faster development and 90% faster model training to deploy a commercial virtual power plant with MathWorks (MATLAB)

VGEN Develops Virtual Power Plant with Deep Learning and Machine Learning

VGEN, a developer of energy management software for virtual power plants (VPPs), faced a tight deadline to deliver commercial EMS that could accurately forecast renewable generation, demand, and market prices and execute constrained storage-and-trade optimizations. Needing faster development than their initial Python approach allowed, VGEN turned to MathWorks tools—principally MATLAB with Deep Learning Toolbox, Optimization Toolbox, Parallel Computing Toolbox, Database Toolbox, and MATLAB Compiler—to accelerate model building and deployment.

Using MathWorks technology, VGEN automated data aggregation and preprocessing, trained hybrid machine learning/deep learning forecasting models, and built an optimization engine and user interface, then packaged the system as standalone apps. The MathWorks-based solution cut development time by about 50% (multiple forecasting and optimization apps delivered in seven months), reduced large‑dataset training time by roughly 90% (10× speedup with datastores and parallel training), and resulted in commercially deployed VPP EMS software in Korea.

VGEN

Seungyup Baek

VGEN