Machine learning in aerodynamic shape optimization

Machine learning (ML) has been increasingly used to aid aerodynamic shape optimization (ASO), thanks the availability of data and continued developments in deep learning. We review applications ML ASO date provide a perspective on state-of-the-art future directions. first introduce conventional current challenges. Next, we fundamentals detail algorithms that have successful ASO. Then, addressing three aspects: compact geometric design space, fast analysis, efficient architecture. In addition providing comprehensive summary research, comment practicality effectiveness developed methods. show how cutting-edge approaches can benefit address challenging demands, such as interactive optimization. Practical large-scale optimizations remain challenge because high cost training. Further research coupling model construction with prior experience knowledge, physics-informed ML, is recommended solve problems.