Return
Transformer-based explicit model predictive control with variable prediction horizon
DOI:10.1016/j.engappai.2026.114345.png)
Abstract
En 中文
• Encoder-only Transformer natively supports variable horizons via parallel inference. • Direct policy optimization eliminates dependence on precomputed optimal solutions. • Outperforms explicit methods in accuracy and speed, vastly faster than online MPC.
Keywords:
Transformer
explicit model predictive control
variable prediction horizon
direct policy optimization
parallel inference
Journal
IF:
8
Papers:
5.3K
Citations:
3.5W

