Unified Energy Flow Model for greenhouse tomato production. Integrates physics-based climate model, mechanistic crop growth (Gompertz sink dynamics), and MPC framework. Based on 6 years of data from a 16,000m² commercial greenhouse in Korea.

Regime-aware quant risk and market stability monitoring framework.

A physics-informed neural network approach is proposed for control-oriented building thermal modeling, combining data with physical laws to improve accuracy and reduce training data needs. The method predicts room temperature, power use, and hidden thermal states more reliably than standard neural networks.