Generative Artificial Intelligence and Computational Design Optimization in Contemporary Architecture: a Systems Engineering Perspective

Abstract: Generative artificial intelligence (GenAI) is increasingly influencing architectural design by expanding computational design exploration, performance evaluation, optimization, and decision-support capabilities. This systematic literature review examines how GenAI and computational design optimization are being integrated into contemporary architecture from a systems engineering perspective. The review followed the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) 2020 reporting framework. Searches were conducted across Scopus, Web of Science, ScienceDirect, SpringerLink, IEEE Xplore, ACM Digital Library, Wiley Online Library, Taylor & Francis Online, SAGE Journals, Google Scholar, and arXiv for literature published between January 2018 and March 2026. The search identified 1,248 records. After removal of 276 duplicates, 972 records were screened by title and abstract, of which 721 were excluded. Full texts were assessed for 251 articles, with 157 excluded for reasons including insufficient architectural relevance, inadequate focus on artificial intelligence, methodological weakness, and unavailable full text. Ninety-four studies were retained for the final synthesis. The review examined computational design, generative AI, multi-objective optimization, systems engineering integration, sustainability, smart buildings, digital twins, and intelligent built environments. The synthesis indicates that GenAI can expand design solution spaces, support rapid concept generation, facilitate performance-based optimization, and strengthen data-driven decision-making. Systems engineering provides a complementary framework through requirements management, lifecycle thinking, feedback mechanisms, risk management, interdisciplinary coordination, and performance evaluation. However, concerns concerning data quality, explainability, interoperability, intellectual property, cybersecurity, professional accountability, and human oversight remain important. The review argues that effective architectural application of GenAI depends not simply on technological adoption but on integrating computational intelligence with systems engineering processes and professional architectural judgment. The reported screening counts establish the review's selection pathway; however, the final manuscript should additionally report the verified Cohen's kappa coefficient and complete study-level MMAT appraisal results from the review records before publication.