.png){kind=icon}
.png){kind=icon}
Return
Multi-fidelity meta-optimization for nature inspired optimization algorithms
.png){kind=icon}
PRE
AI
DOI：10.1016/j.asoc.2020.106619.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
The last couple of decades have witnessed a steadily increasing applications of nature inspired optimization (NIO) in vast fields such as power engineering, environmental engineering, and civil engineering. Behavioral parameters deeply affect the optimization performance for a NIO algorithm. Meta-optimization is good choice for parameter optimization of NIOs, which uses an optimizer to optimize another optimizer. However, meta-optimization is a very time-consuming process. For this reason, we propose a multi-fidelity strategy based meta-optimization approach to speed up the parameter optimization. Four types of fidelity control functions determine the fidelity level in the course of meta-optimization. We test the proposed method in the meta-optimization systems with diverse meta-NIOs (cuckoo search, fruit fly optimizer, gray wolf optimizer, krill herd, and whale optimization algorithm), diverse optimized-NIOs (cuckoo search, differential evolution, particle swarm optimizer, squirrel search algorithm, and water wave optimizer), and diverse benchmark problems (Ackley-50, Eggholder-2, Michalewicz-5, Shubert-2, Sphere-50, and F1-20). We also apply it to a realworld engineering problem to estimate the source terms of gas emission. Experimental results indicate that multi-fidelity strategy can substantially speed up meta-optimization systems and hence has the potential to be generalized to various NIOs. (C) 2020 Elsevier B.V. All rights reserved.
Keywords:
Nature inspired optimization
Meta-heuristics
Parameter optimization
Meta-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:
6.6
Papers:
1.4W
Citations:
4.8W