Accelerating multi-energy system online optimization via integer state variable prediction with operation strategy learning

doi:10.1016/j.energy.2025.139337

Accelerating multi-energy system online optimization via integer state variable prediction with operation strategy learning

2025-11-17

0

PRE

AI

郑成航 (Chenghang Zheng) *

高

高翔 (Xiang Gao)

DOI：10.1016/j.energy.2025.139337

Original

Original request for help

Save

Abstract

En 中文

• Proposes a predictive variable-fixing framework to accelerate real-time multi-energy system optimization. • Achieves over 94.1 % accuracy in state prediction using LSTM-based learning. • Reduces MILP solving time by over 90 % and accelerates solver convergence by 30.1 % using prediction guided methods. • Enhances online optimization responsiveness and improves overall system operation performance.

Accelerating multi-energy system online optimization via integer state variable prediction with operation strategy learning

Abstract

Journal

Organization