Heat transfer optimization in a fluid-filled porous hexagonal cavity using finite element modeling, artificial neural networks, and response surface methodology

Safiya M. Alkindi

doi:10.1016/j.tsep.2026.104871

Surrogate-assisted deep reinforcement learning for intelligent structure design of pre-stressed underground bundles

2026-05-08

0

PRE

AI

S

Shanshan Shen

W

Wen He

Y

Yue Pan

J

Jinghai Gong *

DOI：10.1016/j.autcon.2026.107009

Original

Original request for help

Save

Abstract

En 中文

• SADRL-UBIT framework is developed to provide an effective and automated solution for optimal UBIT structures design. • It integrates Att-Bayes-ANN surrogate model and DQN-based optimization method. • Systematic numerical experiments validate the effectiveness and practicability of the proposed SADRL-UBIT framework.

Keywords:

SADRL-UBIT

surrogate model

deep reinforcement learning

optimal design

pre-stressed underground bundles

Surrogate-assisted deep reinforcement learning for intelligent structure design of pre-stressed underground bundles

Abstract

Journal

Organization

Cited Papers

Citing Papers