.png){kind=icon}
.png){kind=icon}
Return
Data-driven teaching-learning-based optimization (DTLBO) framework for expensive engineering problems
.png){kind=icon}
PRE
AI
DOI：10.1007/s00158-021-03010-1.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
Teaching-learning-based optimization (TLBO) has become a popular heuristic algorithm recently because of its non-parametric characteristic and strong optimization ability. In the paper, the TLBO is improved to solve expensive engineering optimization problems by making full use of the historical process data. An online data-driven learning strategy is adopted in the DTLBO framework, and the surrogate model is adaptively updated with the increase of historical data. The teaching phase of TLBO is modified by predicting teacher and selecting the offspring individuals with the assistance of the constructed surrogate model. The optimization effect and efficiency of the proposed DTLBO algorithm are verified by a series of numerical examples. Moreover, the DTLBO algorithm is applied to solve a typical expensive engineering optimization problem-aerodynamic shape optimization design. The results demonstrate that the proposed DTLBO has the advantage of high efficiency, strong optimization ability, and non-parametric characteristic for expensive engineering problems.
Keywords:
Data-driven
Surrogate model
TLBO algorithm
Non-parametric characteristic
Expensive engineering optimization
AI Summary
Key information extracted from the uploaded paper, including a brief overview, abstract, background, key highlights, visual analysis, and future outlook.
Journal
IF:
4
Papers:
4.8K
Citations:
1.7W