A workflow for climate-adaptive tree planting design to optimize outdoor thermal environment through evolutionary algorithm

Urban areas are increasingly exposed to heat stress, underscoring the need for adaptive, performance-oriented green infrastructure. Tree shading and canopy dynamics play a crucial role in regulating urban microclimates and improving outdoor thermal comfort. This study aims to develop a generative planting design workflow that optimizes tree configurations across seasons. It integrates vegetation parameters, particularly leaf area index (LAI), into a radiation model and explores trade-offs between thermal comfort and the initial purchase cost of trees. The workflow combines Grasshopper and Ladybug Tools with an evolutionary algorithm as an optimizer to generate planting layouts. Field measurements of mean radiant temperature (MRT) were used to validate solar radiation simulations before optimization. The validation improved prediction accuracy in shaded areas by approximately 3 °C. The evolutionary algorithm search revealed multiple planting configurations rather than a single optimal layout. Seasonal differences showed that higher LAI and denser canopies were more effective at summer cooling, whereas sparser structures enhanced winter comfort. Conical-shaped deciduous trees (height = 12 m; trunk diameter = 30 cm) emerged as optimal for both seasons, with LAI values decreasing from 4.6 to 0.6 or 2.8 to 0.0, reflecting seasonal foliage variation. A spacing of 6–12 m further supported adaptability across seasonal contexts. The framework integrates an LAI-revised thermal comfort model with a multi-objective NSGA-II genetic algorithm to generate cost-constrained, seasonally adaptive planting schemes that optimize UTCI across summer and winter conditions. The findings provide guidance for climate-responsive, seasonally adaptive tree-planting strategies in urban landscapes.