Metropolitan service-placement optimization is computationally challenging under strict evaluation budgets and regulatory constraints. Existing approaches either neglect capacity constraints, producing infeasible solutions, or employ population-based metaheuristics requiring hundreds of evaluations—beyond typical municipal planning resources. We introduce a two-stage optimization framework combining Bayesian optimization with domain-informed heuristics to address this constrained, mixed discrete–continuous problem. Stage 1 optimizes continuous service area allocations via the Tree-structured Parzen Estimator with empirical gradient prioritization, reducing effective dimensionality from 81 services to 10–15 per iteration. Stage 2 converts allocations into discrete unit placements via efficiency-ranked bin packing, ensuring regulatory compliance. Evaluation across 35 benchmarks on Saint Petersburg, Russia (117–3060 decision variables), demonstrates that our method achieves 99.4% of the global optimum under a 50-evaluation budget, outperforming BIPOP-CMA-ES (98.4%), PURE-TPE (97.1%), and NSGA-II (96.5%). Optimized configurations improve equity (Gini coefficient of 0.318 → 0.241) while maintaining computational feasibility (2.7 h for 109-block districts). Open-source implementation supports reproducibility and facilitates adoption in metropolitan planning practice.
The article deals with the problem of automation and decision-making support in forming sustainable assortments of plant species for landscaped areas of cities, taking into account environmental factors, conditions of growing, properties of plant species, and landscaping practices adopted in the region. The automation of plant species and composition planning for planted territories requires formation of a knowledge base and output algorithms, as well as consideration of specific landscaping practices and techniques adopted in the region. It also requires the application of new modeling methods in combination with methods of spatial analysis, which together with the database of available regional plant species forms a recommendation system and provides decision support in the tasks of landscaping for city planners. The feature of the proposed method based on network analysis is the combination and composition of plant species, as well as the properties of invasive species, which require the use of graph models to form a knowledge base. The proposed method can be useful for forming sustainable assortments of plant species for landscaping squares, boulevards, small courtyards in the city, technical areas and other cases that do not involve the creation of individual complex landscape design. As part of the testing of the proposed method, the authors have formed a unique information base of species. The use of such a database, together with the method of selecting the species composition based on their resistances and interspecies compatibility will ensure the expansion of the diversity of species in the created ecosystem, which will make green areas more resistant to external anthropogenic factors and diseases, as well as preventing the spread of species invasive to the region.