Optimizing Built Environment Adaptation to Inundation using Genetic Algorithms: Alexandria as a Case Study

Abstract As sea levels rise and extreme rainfall events worsen, coastal cities need to transition from traditional "gray" (fixed) infrastructure towards dynamic, adaptable built environments. Our research focuses on the city of Alexandria, Egypt, to demonstrate how our new computational framework combines high fidelity environmental simulations with Multi-Objective Genetic Algorithms (MOGA) in order to optimize architectural adaptations. We developed a methodology that forms a simulation-optimization loop using hydrostatic & hydrodynamic stress analysis on a digital twin of a residential prototype to assess how a Genetic Algorithm (GA) could evolve different design variables (i.e., floor elevations & porosities) to provide the least structurally vulnerable & most cost-effective designs while still being part of a "Pareto Optimal Front" (POF). The optimized design demonstrated substantial gains in performance by producing a 1.2 m height structure with a 65% porosity on the ground floor, as compared to the base case. It had a 77% lower peak structural stress than the baseline and a restoration cost savings of 65% compared to the baseline. In addition, durability tests (long-term simulation) indicate a 71% reduction in chloride penetration depth, reducing the risk of saltwater corrosion over 20 years. The results of this study support the idea that by implementing AI-enabled generative design, it is possible to develop a scalable process for designing resilient coastal communities and to address the limitations associated with traditional urban planning methods. Lastly, this study provides several recommendations for policymakers about how to incorporate "wet floodproofing" and Dynamic Decision Support Systems into local building regulations.