The application of artificial intelligence (AI) models in maritime and coastal engineering has gained increasing relevance, demonstrating performance comparable to traditional approaches in wave climate analysis and propagation. However, their use in climate change impact and adaptation studies remains limited, particularly for the design and upgrading of coastal protection structures. To address this gap, this study focuses on the development of an AI-based framework to support the adaptation of breakwaters to future climate conditions. A hybrid approach combining artificial neural networks (ANNs) and genetic algorithms (GAs) was implemented, with two feedforward neural networks-based models developed and applied to different sections of the north breakwater of the Port of Valencia, specifically a vertical section and a compound breakwater. The results indicate that, under future climate scenarios (2050), increases of up to 1.2 m in crest elevation, together with reinforcement of the armor layer, are required to ensure adequate structural performance. The analysis also highlights the critical role of extreme events, as approximately 60% of the model errors were concentrated in the upper 90th percentile of wave conditions. Overall, the proposed hybrid ANN-GA framework demonstrated very strong performance, achieving computational efficiencies 30 to 40 times greater than ANN-only models in terms of computational time. These findings underscore the necessity of adapting coastal structures to climate change and confirm the potential of AI-based models as effective tools for climate-resilient coastal engineering, while emphasizing the importance of accurately representing extreme wave conditions.
Juan et al. (Fri,) studied this question.