Research on the Method of Automatic Generation and Multi-Objective Optimization of Block Spatial Form Based on Thermal Comfort Demand

Abstract

Urban thermal environment challenges in China have made outdoor thermal comfort a key factor in evaluating spatial quality and livability. Building layout not only affects internal performance but also shapes the microclimate of surrounding outdoor spaces. The climatic characteristics of temperate monsoon climate regions significantly impact residents’ outdoor activities. Most existing studies focus solely on either the external thermal environment or the buildings themselves in isolation. This study focuses on Beijing, a representative city in the temperate monsoon climate zone, and explores block-scale spatial optimization using computational typology. The objective is to balance architectural performance with outdoor thermal comfort in both winter and summer. Optimization targets include the Universal Thermal Climate Index (UTCI), winter sunshine duration, and summer solar radiation. Results show winter UTCI can be optimized to −6.13 °C to −1.18 °C and summer UTCI to 28.19 °C to 29.17 °C, with greater optimization potential in winter (23.5% higher). Synergistic relationships are observed between winter comfort and sunshine duration (coefficient: 0.777) and between summer comfort and solar radiation (coefficient: 0.947). However, trade-offs exist between seasonal comfort indicators, with strong conflicts between winter and summer objectives. Two distinct form types—“low-south-high-north enclosed” for winter and “high-rise point-type low-density” for summer—are identified as effective for seasonal adaptation. The study proposes an integrated method combining data-driven generation, multi-objective optimization, and clustering-based decision-making. This approach moves beyond traditional empirical design, offering a quantitative and adaptable strategy for climate-responsive urban block planning and supporting low-carbon urban transformation.

1. Introduction

1.1. Research Background

The rapid urbanization process has significantly altered the thermal dynamics of urban environments, with the problem of the urban heat island effect becoming increasingly prominent [1]. In areas with high population density, the superimposed effect of rising temperatures and poor air circulation further intensifies the heat island intensity. This not only increases environmental risks but also directly threatens the health and safety of residents [2,3]. A large body of research has confirmed a strong correlation between urban morphological elements such as building density and layout, and road network structure, and the microclimatic environment [4,5]. Optimizing spatial configuration to mitigate thermal environmental risks has become a consensus in academic circles.

In China, the annual urban built-up area expansion rate of 1.36% (National Bureau of Statistics, 2023) has spawned a high-density urban pattern [6], leading to an increase in the summer heat island intensity of 32 key cities at a rate of 0.12 °C per decade [7]. In areas with a building density exceeding 50%, 35% of plots suffer from less than 1 h of winter sunlight [1]. This development model of “prioritizing development over climate adaptation” has caused a continuous deterioration of the urban microclimate, with residents’ heat stress risks increasing by 47% compared to ten years ago [2], exposing the deep-seated contradiction between urban morphology and climate adaptability. Although the coupling relationship between morphological elements such as building density and height gradients and indicators such as UTCI (Universal Thermal Climate Index) and solar radiation has been confirmed [4,5], traditional planning relies on static control indicators (such as the sunshine spacing coefficient D/H = 1.2), which struggle to balance the differentiated climate needs between winter and summer. For example, while the row-type layout commonly adopted in South China can improve summer ventilation efficiency by 15%, it also results in a 20% increase in heat loss during the winter months [8].

To bridge this mismatch, recent studies have advocated for adaptive urban morphology strategies that dynamically integrate environmental responsiveness into spatial planning. As Sádaba et al., (2025) demonstrate, embedding nature-based solutions into urban infrastructure design can significantly enhance climate adaptability by addressing seasonal thermal variations and reinforcing ecological resilience in compact urban fabrics [9]. Their case study illustrates how urban form, when flexibly adapted to climatic demands, can transcend the limitations of static regulatory tools and mitigate both overheating and winter heat loss synergistically.

The development of digital technologies has provided a new paradigm for solving this challenge. The deep integration of the Rhino + Grasshopper parametric platform with simulation tools such as ENVI-met and Ladybug has made data-driven morphological optimization feasible. Liu et al., (2023) achieved synergistic improvements in daylighting, energy consumption, and photovoltaic potential through a multi-objective optimization framework [10]. However, existing research still has two major limitations: first, it focuses on single-climate objectives (such as summer thermal comfort or winter sunlight), lacking a quantitative analysis of the multi-objective coordination–conflict mechanisms between winter and summer; second, the processes of form generation and optimization are fragmented, and an automated closed loop of “parametric modeling–multi-objective optimization–rule extraction” has not yet been established, with the optimization cycle for a single scheme generally exceeding 20 h.

Against the backdrop of the imbalance between urban morphology and microclimate in China’s urbanization process, this study addresses the climatic characteristics of long annual discomfort periods in temperate monsoon climate regions. It aims to develop an automated method for generating and optimizing block spatial forms that balances both architectural performance and external environmental quality. Based on the results of multi-objective optimization, the study further extracts spatial improvement strategies tailored to the specific characteristics of the target area. By quantitatively analyzing the dynamic interaction laws between urban morphological parameters and multi-climate objectives (e.g., thermal comfort, sunlight, and radiation in winter and summer), the research reveals their coordination–conflict mechanisms, filling the gap in quantitative studies on seasonal climate adaptation strategies. Additionally, by integrating parametric modeling, multi-physical-field simulation, and genetic algorithms into an automated toolchain, the study promotes the transformation of urban design from “empirical trial-and-error” to “data-driven,” providing scientific support and practical pathways for constructing low-carbon and healthy urban microclimatic environments.

1.2. Previous Studies

1.2.1. Advances in Thermal Comfort Assessment and Simulation Techniques

Outdoor thermal comfort, as a core indicator of urban livability, has become a central focus in academic research. According to ASHRAE, thermal comfort is defined as “the state of mind that expresses satisfaction with the thermal environment [11].” Current research on the relationship between urban morphology and thermal performance can be broadly categorized into two dimensions.

The first dimension involves the Urban Thermal Environment Index (UTEI), which encompasses indicators such as solar radiation [12], surface temperature [13], surface urban heat island intensity [14], and overall urban heat island intensity [3]. These parameters are essential for analyzing urban heat island effects and their implications for building energy consumption [15].

The second dimension is the Personal Thermal Comfort Index (PTCI), which evaluates human thermal stress by integrating key environmental parameters such as air temperature, thermal radiation, wind speed, and humidity into equivalent temperature indices—such as Standard Effective Temperature (SET), Physiological Equivalent Temperature (PET), and Universal Thermal Climate Index (UTCI) [16]. Research has demonstrated that morphological elements—including building height, footprint area, and green space configuration—have a significant impact on the thermal environment. For instance, solar radiation in urban canyons with high green coverage can be up to four times lower than in open streets [17]. In Shenzhen, a combination of vertical greening, high-albedo paving materials, and ground vegetation has successfully maintained UTCI within comfort thresholds [18]. However, most existing studies tend to focus on either UTEI or PTCI in isolation, resulting in a fragmented understanding of urban thermal dynamics. There is an urgent need for an integrated approach that combines thermal performance metrics with spatial morphological characteristics to enhance the accuracy of urban planning and design while addressing growing climate-related risks [19].

At the technical tool level, the Rhino + Grasshopper parametric platform has established a mature optimization toolchain. Galapagos enables single-objective optimization, Octopus supports iterative multi-objective optimization, and Wallacei enhances multi-objective decision-making through integrated analysis and visualization features [20]. These tools have been widely used in building performance optimization tasks, including energy efficiency improvements and thermal comfort balancing, particularly in retrofit scenarios. For example, Li et al., (2021) demonstrated the effectiveness of optimization frameworks in achieving dual performance goals of energy savings and occupant comfort in green retrofit projects, confirming the value of parametric platforms in real-world applications [21]. Similarly, in the Jianhu case study, a block morphology optimization framework developed using Grasshopper and Wallacei X (https://www.wallacei.com), successfully achieved synergistic improvements in both energy efficiency and environmental performance, demonstrating the feasibility of data-driven methods for optimizing complex urban systems [10].

1.2.2. Algorithm Optimization and Multi-Objective Integration

Urban morphology optimization has become a key strategy for enhancing environmental performance. Early studies primarily focused on the building scale or single-performance objectives, such as reducing energy consumption or improving thermal comfort [22]. With the advancement of simulation tools like ENVI-met and Ladybug, the scope of optimization has expanded to the block scale, enabling multi-objective coordination [23]. This includes diverse environmental goals such as solar radiation regulation [24], enhancement of outdoor thermal comfort [25], and optimization of natural ventilation [26].

At the algorithmic application level, genetic algorithms (GAs) have become the mainstream choice due to their strong global optimization capabilities. In practical applications, Shan et al. (2023) employed single-objective optimization to analyze the impact of façade morphology on the thermal environment [27]; Matallah et al. (2021) integrated the Grasshopper platform with GA to predict thermal comfort under extreme climatic conditions [28]; in China, Yin Chenhuan (2020) utilized the Galapagos plugin to optimize summer thermal conditions in arid–hot regions [29].

Despite these fruitful achievements, existing research still has significant limitations: First, there is a fragmented analysis between building performance and the outdoor environment. Most studies discuss building performance (such as sunlight and energy consumption) or outdoor thermal environments (UTCI) in isolation, ignoring their coupling effects. This fragmentation makes it impossible to fully analyze the overall impact mechanism of urban morphology on the microclimate.

Second, there is a one-sided approach to multi-objective optimization. Existing research mainly focuses on single-objective optimization, and even when involving multiple objectives, it is mostly limited to internal systems (such as building energy consumption and daylighting), lacking cross-system balance between building performance and outdoor environments. Additionally, there is a lack of comprehensive automated frameworks that integrate 3D urban form generation with multi-objective optimization based on different planning and design requirements.

Third, there is a shortage of research on temperate monsoon climate regions. Most existing achievements focus on extreme climate regions such as severe cold and hot–humid areas, with insufficient research on adaptive strategies for temperate monsoon climate regions (covering approximately 22.2% of China’s land area). The distinct seasonal characteristics of cold winters and hot summers in this climate zone require designs to balance cross-seasonal performance.

In response to the shortcomings of previous research, the innovations of this study lie in systematically addressing the core gaps in the existing literature—most studies either analyze outdoor thermal environments in isolation or only focus on building performance, while neglecting the synergistic and conflicting relationships between the two.

To address this gap, this study, for the first time, integrates outdoor thermal comfort (UTCI) with building performance indicators (sunlight duration, solar radiation) into a unified multi-objective optimization framework. Quantifying the dynamic interactions between urban morphological parameters (e.g., building types, block road widths) and climate objectives across winter and summer, it reveals the synergistic–conflicting mechanisms between outdoor comfort and building performance.

This approach breaks through the traditional “single-system” research paradigm and establishes a data-driven climate-adaptive design method capable of balancing annual environmental demands. For temperate monsoon climate regions, the study further proposes season-specific morphological strategies, providing a quantifiable climate-resilient design framework that is absent in the existing literature for this climate zone.

2. Methods

The research process is divided into two main parts: urban block analysis and the development of a parametric platform, as illustrated in Figure 1. In the urban block analysis phase, the study first examines urban blocks of Beijing, China based on existing issues to identify typical morphological patterns and scales. Since streets and buildings are key components of urban block morphology, they directly influence building performance and the external environment. Based on this, the study classifies building cluster prototypes and road prototypes that align with the climatic characteristics of the region. These prototypes serve as standard units for constructing typical urban blocks. Relevant performance indicators are then quantified, forming the basis for the automated generation of block layouts [30,31,32].

For the development of the optimization platform, this study utilizes Rhino + Grasshopper as the parametric design platform, which consists of five main modules: parameter presetting, block generation, performance simulation, algorithmic optimization, and data recording and processing [33,34]. The platform controls key morphological parameters of urban blocks through Rhino and Grasshopper, while environmental performance simulation tools are employed to evaluate target performance indicators. Finally, a multi-objective optimization process uses evolutionary algorithms to generate the optimal climate-responsive urban morphology. The results are recorded and quantitatively analyzed to inform design improvements [35,36].

2.1. Extraction of Typical Urban Blocks and Building Prototypes

2.1.1. Typical Urban Block Morphology and Scale

The study area is located in the Haidian District, Beijing, within a temperate, humid monsoon climate zone. The Universal Thermal Climate Index (UTCI) averages −8.1 °C in winter and 32.0 °C in summer, with a thermal discomfort period accounting for 67% of the year (Figure 2).

As shown in Figure 3, the study area covers 17 km², characterized by a high population density and a diverse range of building types, making it well-suited for morphological classification. The outdoor spaces within this area include urban arterial roads, plazas, and green parks.

A survey was conducted on eight urban blocks within the study area, with relevant information summarized in

Table 1. Morphological information of case blocks.

Surveyed Case Number	A	B	C	D
Block Morphological Layout
Block Scale (Length × Width)	424 m × 363 m	241 m × 294 m	340 m × 379 m	370 m × 265 m
Number of Clusters	11	6	8	8
Average Size of Clusters	125 m × 109 m	119 m × 88 m	112 m × 130 m	103 m × 118 m
Road Morphology (Road Width)	Grid Type (6–20 m)	Grid Type (8–25 m)	Grid Type (10–20 m)	Grid Type (7–25 m)
Surveyed Case Serial Number	E	F	G	H
Block Morphology Layout
Block Scale (Length × Width)	249 m × 366 m	380 m × 260 m	376 m × 347 m	295 m × 305 m
Number of Clusters	11	6	7	6
Cluster Scale (Length × Width)	76 m × 102 m	140 m × 114 m	138 m × 154 m	123 m × 114 m
Road Morphology (Road Width)	Grid Type (10–20 m)	Grid Type (8–25 m)	Grid Type (10–30 m)	Grid Type (10–28 m)

. Block shapes are primarily rectangular and oriented in north–south or east–west directions. The area of each case block ranges from 70,854 m² to 153,912 m², with an average of 107,757 m². For the north–south length, the maximum value is 424 m, the minimum is 241 m, and the average is 334 m; for the east–west length, the minimum is 260 m, the maximum is 379 m, and the average is 322 m. The number of internal clusters in each block ranges from 6 to 11, with an average cluster size of 107 m × 106 m. All blocks adopt a grid layout regarding road morphology, and road widths mostly fall within the 8–22 m range.

Based on the above research on block morphology and dimensions, this paper takes a 100 m × 100 m basic unit as the plot for clusters, adopts a 3 × 3 grid as the inter-cluster road network structure within the block, and sets road widths as two types (10 m and 20 m) according to actual road widths. A rectangular abstract typical block model of 320 m × 320 m or 340 m × 340 m is formed. Relevant data are detailed in

Table 2. Comparison between actual blocks and abstract typical blocks.

		Information on Actual Blocks		Information on the Abstracted Typical Block Model
Block Morphological Layout Diagram
Data of the Overall Block	Form	Rectangle		Rectangle
	Length in the North–South Direction	An average of 334 m.		320 m/340 m
	Length in the East–West Direction	An average of 322 m.		320 m/340 m
	Extent	An average of 107,757 m2		102,400 m2/115,600 m2
Information on Internal Clusters	The number of clusters	From 6 to 11		9
	The size of the clusters	An average of 107 m × 109 m		100 m × 100 m
Information about the Road	Form	Grid Type		Grid Type
	Width	8–22 m		10 m/20 m

.

2.1.2. Extraction of Building Prototypes

Due to the influence of natural conditions such as geographical location and climatic conditions, cities in different regions develop specific architectural forms. Representing urban and architectural forms graphically is the most direct expression method in architecture, serving as the foundation for parametric research on physical forms and constituting the core issue of parametrization [17]. In the 1970s, the Martin Centre for Architectural and Urban Studies at the University of Cambridge summarized the geometric morphological characteristics of modern European cities and buildings [37] and proposed basic prototypes of urban morphology. To date, the fundamental models used in Western studies of urban microclimatic environments still employ the urban morphological prototypes developed by the Martin Centre.

This study summarizes the architectural form characteristics of the research area. It extracts urban form prototypes by utilizing the data from open-platform sources such as OpenStreetMap (OSM) and Baidu Maps [38,39], along with on-site investigations.

Table 3. Urban Form Prototypes in Haidian District, Beijing.

Serial Number	Typical Model	Plane	Source of Type	Relevant Parameters
Type 1				Plane Type: Point-type Plot Ratio: 0.58–0.86 Building Density: 0.29 Number of Building Floors: 2–3 Building Height: 6–9 m
Type 2				Plane Type: Point-type Plot Ratio: 2.7–5.5 Building Density: 0.17 Number of Building Floors: 16–33 Building Height: 50–100 m
Type 3				Plane Type: Point-type Plot Ratio: 2.6–5.3 Building Density: 0.16 Number of Building Floors: 16–33 Building Height: 50–100 m
Type 4				Plane Type: Slab-type Plot Ratio: 1.1–1.3 Building Density: 0.22 Number of Building Floors: 5–6 Building Height: 15–20 m
Type 5				Plane Type: Slab-type Plot Ratio: 2.6–3.2 Building Density: 0.17 Number of Building Floors: 16–20 Building Height: 48–60 m
Type 6				Plane Type: Enclosed-type Plot Ratio: 1.9–3.0 Building Density: 0.38 Number of Building Floors: 5–8 Building Height: 15–24 m

shows six prototypes extracted and classified into three major categories: point-type, slab-type, and enclosed-type. The heights of building types are divided into three ranges: low-rise, multi-story, and high-rise.

2.2. Platform Construction

2.2.1. Parameter Presetting and Model Generation

The parameter selection in this study follows the principles of “goal-oriented, climate-adaptive, and data-driven,” with specific scientific justifications as follows:

(1) Optimization objectives are selected with climate theory and regulatory requirements as the core

The selection of winter and summer UTCI, sunshine duration, and solar radiation is supported by clear theoretical foundations. As a thermal comfort assessment model recommended by ISO [40] UTCI integrates multi-physical field parameters such as air temperature, radiation, wind speed, and humidity, directly reflecting the human thermal stress state. This aligns with the characteristics of temperate monsoon climate regions (average winter UTCI of −8.1 °C and summer UTCI of 32.0 °C). Sunshine duration (≥2 h on the winter solstice) and solar radiation are based on standards such as the Urban Residential Area Planning and Design Standards [41], forming a dual-dimensional assessment system of “human comfort-building energy efficiency.”

(2) Independent variables (morphological parameters) are screened based on the theoretical mechanisms and data validation

The determination of building cluster types and block road widths integrates theoretical derivation and empirical analysis: the impacts of enclosed, point-type, and slab-type layouts on winter–summer wind environments and sunlight are rooted in theoretical achievements from institutions such as the Martin Centre at the University of Cambridge. Block road width design (10 m/20 m) aligns with the “street canyon effect” theory [42], where narrow roads enhance winter enclosure while wide roads promote summer ventilation. The influence of 12 parameters was quantified through pre-experimental ENVI-met simulations, and six prototype types were abstracted from eight typical blocks in Beijing (Figure 4).

(3) The parameter system achieves scientific consistency through a closed loop of ‘theory–data–algorithm’

The study establishes a linkage model between “morphological parameters–climate objectives,” validates parameter sensitivity through the Grasshopper + Wallacei platform and NSGA-II algorithm, and finally generates Pareto solutions. The overall logical chain is grounded in climate theory and supported by algorithmic optimization, demonstrating that the system is rational and effective for multi-objective optimization.

2.2.2. Performance Simulation

Based on the above optimization objectives, performance simulations are primarily divided into three components: outdoor wind environment simulation, outdoor thermal environment simulation, and building solar radiation simulation.

First, the wind environment simulation is achieved through the combination of Eddy 3D [43] and Open FOAM. First, the Grasshopper parametric model is exported in STL format, imported into Eddy 3D to generate computational grids (minimum grid size 0.5 m, near-ground area encrypted to 0.2 m), and the Open FOAM solver is called for CFD simulation. The boundary conditions are set as follows: a winter wind speed of 2.7 m/s, wind direction 15° west of north (corresponding to the 22% dominant wind frequency in Beijing’s winter in the China Building Thermal Environment Analysis Special Meteorological Dataset); a summer wind speed of 2.2 m/s, wind direction due south (matching the characteristics of summer southeast monsoons); and the k-ε turbulence model is used, considering the building surface roughness (roughness length 0.01 m).

Second, the thermal environment and solar radiation simulation rely on the Ladybug Tools and Radiance platforms. The outdoor thermal comfort assessment is completed through Ladybug’s UTCI Calculator component, inputting EPW meteorological data, human body parameters (winter clothing quantity 1.5clo, summer 0.5clo), and 1.5 m measuring point height data, coupling the wind speed field output from the wind simulation and the mean radiant temperature (MRT) calculated by Radiance. The sunshine duration is calculated by the Daysim module for the proportion of areas with cumulative sunshine ≥2 h on the winter solstice (December 22) (grid accuracy 1 m × 1 m); the summer solar radiation is simulated by Honeybee calling the Radiance engine for the total horizontal radiation of a typical week (24–30 August), with a time step of 1 h, and the model includes building shading effects and surface albedo parameters (default 0.3).

Finally, in terms of meteorological data and simulation cycles, the EPW file for the Beijing area (ID: CN_BJ_Beijing.545110_IWEC) downloaded from the Energy Plus [44] official website is used, with data covering hourly meteorological information from 1981 to 2010 (such as summer average humidity 75%, winter 40%). The simulation time is set as follows: the typical cold week (12.22–12.28) is selected for winter, covering the low-temperature extreme values around the winter solstice; the typical hot week (8.24–8.30) is selected for summer, corresponding to the high-temperature period of the dog days.

2.2.3. Algorithm-Based Optimization Search

This study selects Wallacei as the optimization processing platform for multi-objective optimization. It internally integrates the NSGA-II multi-objective optimization algorithm, an evolutionary engine that performs well in optimizing more than three objective values [45]. In Wallacei, since the default setting of the plugin is to optimize toward the minimum value, the outdoor comfort (UTCI) in winter and the building duration target on the winter solstice should be maximized. Therefore, negative values should be taken for both as optimization index values. The remaining two objectives retain their original values as index values. The four optimization objectives are then connected to the Wallacei plugin for optimization. Relevant parameter settings include a gene mutation rate of 0.2, a price difference probability of 0.9, and both crossover and mutation distribution indices of 20. Due to the long wind simulation time and constraints on iteration time, the number of operational generations is set to 50, with each population size set to 30. The construction process of the optimization platform is shown in Figure 5.

3. Results

3.1. Convergence Analysis

After nearly 350 h of iterative calculation, 1500 solutions were obtained, and the overall state of the population tended to be stable. As shown in Figure 6, the standard deviation distribution maps of each optimization objective reflect the performance fluctuation degree of different design schemes in terms of winter/summer UTCI and sunshine indicators. A smaller standard deviation indicates that the target values are more concentrated and the design stability is higher. These maps can be used to evaluate the robustness of optimization results and help screen design schemes with balanced performance. The trend of the standard deviation of each objective gradually flattened out, indicating that the optimization results of each objective have a convergent trend. Due to the vast model search space and the time limit for wind environment optimization, although complete convergence has not been achieved, many Pareto front solutions can still be screened out to a great extent.

As shown in Figure 7, the number of solutions generated by the optimization increases with the number of iterations. There are 277 solutions globally, which ensures the diversity of the optimization solutions and proves that the optimization experiment can provide more solutions with better performance for designers.

The distribution of the index values of each optimization objective is shown in Figure 8. The distribution of each optimization objective’s index values is shown in Figure 8. The multi-axis spatial index value distribution map intuitively displays the distribution of different design schemes across key performance indicators such as winter thermal comfort, summer thermal comfort, sunshine duration, and solar radiation. This map reveals the mutual constraints between objectives through a multi-axis coordinate system, helping to identify Pareto frontier solutions and providing decision support for building performance optimization design. In terms of outdoor thermal comfort in winter, the variation range of its index values is 1.18 to 6.13; that is, the range of the winter UTCI value is −6.13 °C to −1.18 °C (since the optimization direction is opposite to the target direction, a negative value is taken). Overall, it is higher than the average winter UTCI value (−8.1 °C) in the Beijing area, and the variation range is relatively large, indicating that this objective has great optimization potential.

The variation range of the index values of outdoor thermal comfort in summer is relatively tiny. The variation range of the summer UTCI is from 28.19 °C to 29.17 °C, which is overall lower than the average summer UTCI value (32 °C) in the Beijing area. In terms of the building sunshine duration on the winter solstice, the variation range of its index values is −4.12 to −3.02; that is, the variation range of the building sunshine duration on the winter solstice in the solution set is from 3.02 h to 4.12 h. The range of the available radiation number of buildings in summer is from 12.07 kWh/m² to 14.49 kWh/m², and each objective has been effectively optimized. Based on the above, the simulation-based optimization experiment has been successful.

3.2. Pareto Front Analysis

The distribution of the complete solution set within the objective performance space is shown in Figure 9a. Among them, the blue-marked points represent the final generation of Pareto-optimal solutions, totaling 30 design schemes. These final-generation solutions were analyzed across multiple dimensions to explore relationships between performance indicators. As illustrated in Figure 9b, a clear synergistic relationship exists between the winter outdoor thermal comfort indicator and the duration of solar exposure on the winter solstice. The correlation coefficient is 0.777, indicating that when the solar exposure performance improves (i.e., lower indicator value and longer sunshine duration), the outdoor thermal comfort in winter also tends to strengthen somewhat. Similarly, as shown in Figure 9e, the summer outdoor thermal comfort indicator exhibits a strong positive correlation with the summer solar radiation received by buildings. The correlation coefficient is 0.947, suggesting that when buildings receive less solar radiation during summer (i.e., lower indicator value and reduced radiation intensity), the summer UTCI is correspondingly lower, indicating enhanced outdoor thermal comfort.

In addition to the observed synergistic relationships, competitive relationships exist among specific objectives, highlighting the significance of employing multi-objective optimization algorithms to enhance individual performance indicators and effectively manage trade-offs between conflicting goals. As illustrated in Figure 9c,d, there is a clear trade-off between winter outdoor thermal comfort and both summer outdoor thermal comfort and summer solar radiation received by buildings. The correlation coefficients are −0.975 and −0.874, respectively. This indicates that when winter outdoor thermal comfort improves (i.e., lower indicator value and higher winter UTCI), summer thermal comfort deteriorates, and buildings tend to receive more solar radiation in summer, reflected in higher indicator values and reduced overall performance. Similarly, competitive relationships are observed between solar exposure duration on the winter solstice and summer outdoor thermal comfort and solar radiation, as shown in Figure 9f,g. The correlation coefficients are −0.871 and −0.973, respectively, suggesting that design strategies that maximize winter solar access often lead to increased summer heat gain and reduced outdoor comfort during hot periods. These results emphasize the importance of identifying an optimal balance among conflicting seasonal performance goals, reinforcing the utility of the Pareto front as a decision-making tool for generating climate-responsive urban forms that are seasonally adaptive and context-sensitive.

3.3. Analysis of Influencing Factors of Each Objective

The building form and block layout directly impact the four optimization objectives: building sunshine duration, radiation distribution, and outdoor comfort levels in winter and summer. These objectives influence and restrict each other. To explore the influencing factors of each objective, the top ten schemes for each purpose are selected from all the schemes as the relatively optimal solution set for a single objective (for example, the top ten schemes with a higher winter UTCI are taken as the relatively optimal solution set for outdoor thermal comfort in winter). The statistical results of each solution set are shown in

Table 4. Statistics of the relevant objective values of the schemes in the relatively optimal solution set for each objective.

Single-Objective Ranking		Relatively Optimal Solution Set for Outdoor Thermal Comfort in Winter	Relatively Optimal Solution Set for Outdoor Thermal Comfort in Summer	Relatively Optimal Solution Set for the Duration of Building Sunlight Exposure on the Winter Solstice	Relatively Optimal Solution Set for the Available Radiation Quantity of Buildings in Summer
		Objective: UTCI in Winter	Objective: UTCI in Summer	Objective: The Duration of Sunlight Exposure on the Winter Solstice	Objective: The Available Radiation of Buildings in Summer
Model of the Optimal Scheme for a Single Objective and the Objective Value
		−1.18 °C	28.19 °C	4.12 h	12.07 kWh/m2
Objective Values of the Schemes Ranked from Second to Tenth	2	−1.23 °C	28.20 °C	4.10 h	12.09 kWh/m2
	3	−1.29 °C	28.20 °C	4.07 h	12.09 kWh/m2
	4	−1.33 °C	28.21 °C	4.06 h	12.12 kWh/m2
	5	−1.35 °C	28.21 °C	4.06 h	12.15 kWh/m2
	6	−1.37 °C	28.22 °C	4.05 h	12.17 kWh/m2
	7	−1.43 °C	28.23 °C	4.05 h	12.17 kWh/m2
	8	−1.47 °C	28.24 °C	4.04 h	12.18 kWh/m2
	9	−1.49 °C	28.26 °C	4.02 h	12.20 kWh/m2
	10	−1.51 °C	28.27 °C	4.00 h	12.21 kWh/m2
Average		−1.45 °C	28.23 °C	4.05 h	12.15 kWh/m2

.

In the relatively optimal solution set for outdoor comfort in winter, the average value of the winter UTCI is −1.45 °C, and the winter UTCI of the optimal scheme in this solution set is −1.18 °C. In the relatively optimal solution for outdoor comfort in summer, the summer UTCI ranges from 28.19 °C to 28.27 °C, with an average value of 28.23 °C. In the relatively optimal solution set for the sunshine duration on the winter solstice, the building sunshine duration on the winter solstice ranges from 4.00 h to 4.12 h, with an average value of 4.05 h. In the relatively optimal solution for solar radiation in summer, the average available radiation of buildings in summer is 12.15 kWh/m².

We classified and statistically analyzed the influence of factors such as the building height, building density, and plot ratio of the overall block, according to each objective’s relatively optimal solution sets (Figure 10). In terms of building height, as shown in Figure 10a, the average building height of the blocks in the relatively optimal solution for outdoor comfort in winter is the lowest, with a value of 30.3 m, followed by the sunshine duration on the winter solstice, with a value of 41.7 m. The building heights in the relatively optimal solution sets for outdoor comfort in summer and solar radiation in summer are significantly higher than the previous two. Among them, the highest building height in the relatively optimal solution for outdoor comfort in summer is 93.7 m. Regarding building density, the density of the relatively optimal solution sets for outdoor comfort in winter and the duration of sunshine on the winter solstice are significantly higher than those in the relatively optimal solution sets for outdoor comfort in summer and solar radiation in summer. Among them, the building density in the solution set for outdoor comfort in winter is the highest, reaching 0.33, and the building density in the solution set for outdoor comfort in summer is the lowest, with a value of 0.16. See Figure 10b for details. In terms of plot ratio, as shown in Figure 10c, the plot ratios in the relatively optimal solution sets for outdoor comfort in winter and the sunshine duration on the winter solstice are significantly lower than those in the relatively optimal solution sets for outdoor comfort in summer and solar radiation in summer. Among them, the plot ratio in the solution set for outdoor comfort in winter is the lowest (with a value of 2.7), and the plot ratio in the solution set for outdoor comfort in summer is the highest, reaching 5.1.

Regarding building types, the planar types are mainly point-type in the relatively optimal solution sets for outdoor comfort and solar radiation in summer. Their proportions in the two solution sets are 84% and 65%, respectively, followed by the slab-type, the highest-rise buildings. In the relatively optimal solution for outdoor comfort in winter, the planar type is mainly the enclosed type (accounting for 67%), primarily low-rise and multi-story buildings. In the relatively optimal solution set for the sunshine duration on the winter solstice, the point-type buildings account for the most significant proportion (accounting for 58%), followed by the enclosed type (accounting for 40%), primarily low-rise and multi-story buildings. See Figure 10d,e for relevant statistics. As shown in Figure 10f, in the relatively optimal solution sets for outdoor comfort in winter and solar radiation in summer, the road width of the blocks is mainly 10 m, accounting for 80% and 70%, respectively. In the relatively optimal solution for outdoor comfort in summer, the road widths of 10 m and 20 m each account for 50%. In the relatively optimal solution for sunshine duration on the winter solstice, the road width is mainly 20 m, accounting for 90%.

When analyzing the influencing factors of each objective, it is necessary to consider the relevant building factor indicators of the overall block and the composition of the building indicators of the groups in different orientations within the block. We numbered the groups according to their various positions (Figure 11) and statistically analyzed the building indicators of the groups at each position (

Table 5. Building indicators of groups at different positions.

	Relatively Optimal Solution Set for Outdoor Thermal Comfort in Winter	Relatively Optimal Solution Set for Daylight Duration on the Winter Solstice	Relatively Optimal Solution Set for Outdoor Thermal Comfort in Summer	Relatively Optimal Solution Set for Building Radiation in Summer
Building height
Building density
Plot ratio

). In the relatively optimal solution for outdoor comfort in winter, the building height increases sequentially from south to north, and the position of No. 5 is the highest (with an average value of 90.1 m). Regarding the layout of building density, the positions of groups No. 4 and No. 5 are relatively low at 0.24 and 0.19, respectively, and the other groups are relatively high (with an average value of 0.35). Overall, it shows a trend of being high around the edges and low in the middle.

In the relatively optimal solution set for the duration of building sunlight exposure on the winter solstice, the layout of building height generally shows a trend of being low in the south, high in the north, low in the east, and high in the west. The No. 2 and No. 5 positions are significantly higher than the others (87 m and 100 m, respectively). Regarding building density, the positions of groups No. 2 and No. 5 are relatively low at 0.18 and 0.16, respectively, and the other groups are relatively high. The plot ratio generally shows a trend of being low in the south and high in the north, and low in the east and high in the west. The positions of No. 0 and No. 4 are relatively low (both are 0.36), and the positions of groups No. 2 and No. 5 are relatively high at 4.73 and 5.36, respectively.

In the relatively optimal solution set for outdoor thermal comfort in summer, the average building height of the group at position No. 6 is relatively low (with a value of 65.7 m), and the difference between the other positions is slight, ranging from 88.6 m to 100 m. In the relatively optimal solution set for the radiation quantity of buildings in summer, the layout of building height generally shows a trend of being high in the middle and low around the edges. Among them, the building heights at positions No. 0, No. 2, No. 6, and No. 8 are relatively low, ranging from 65.7 m to 71.4 m.

4. Clustering Analysis and Scheme Screening

The simulation calculations ultimately yielded a total of 277 global Pareto frontier solutions, with their three-dimensional spatial distribution of objective performances shown in Figure 12a. To rapidly classify the frontier solutions and extract representative block morphologies, cluster analysis was performed on the frontier solutions using the K-means algorithm. The clustering principle was to ensure that the differences in objective performance between different groups after clustering were maximized, while the performance of solutions within the same group was minimized. Through repeated trials and comparisons, it was found that when the clustering number K was set to 6, the classification effect better satisfied the above principles. As shown in Figure 12b, six groups were obtained after clustering, with group numbers Cluster 1 to Cluster 6.

The performance of each objective in different groups is shown in Figure 13. Regarding outdoor thermal comfort in winter, Clusters 1 and 5 have better optimization effects. The average values of winter UTCI values of the two groups are relatively high, which are −1.95 °C and −2.18 °C, respectively. Cluster 2 has a poor optimization effect, with an average value of −5.17 °C. Regarding outdoor thermal comfort in summer, Clusters 2 and 6 have better optimization effects. The average values of summer UTCI of the two groups are relatively low, at 28.2 °C and 28.46 °C, respectively. Cluster 1 has a poor optimization effect. Regarding the duration of building sunlight exposure on the winter solstice, Cluster 1 and Cluster 3 have better optimization effects. The average values of sunlight exposure duration of the two groups are relatively high, at 3.93 h and 3.79 h, respectively. Cluster 2 has a poor optimization effect, and the average value of sunlight exposure duration of this group is 3.20 h (note that here, the unit in the original text “3.20 °C” should be “3.20 h”, which is a wrong unit). Regarding the solar radiation of buildings in summer, Cluster 2 has the best optimization effect, with a relatively low available number of buildings, and the radiation amount per unit area is 12.31 kWh/m². Cluster 1 has a poor optimization effect, and this group’s radiation amount per unit area reaches 14.19 kWh/m².

Overall, Cluster 1 has a better optimization effect in terms of outdoor thermal comfort in winter and the duration of building sunlight exposure on the winter solstice, but has a poor optimization effect in terms of outdoor thermal comfort in summer and the available radiation of buildings in summer. Cluster 2, on the other hand, has a better optimization effect in terms of outdoor thermal comfort in summer and the available radiation of buildings in summer, but has a poor optimization effect in terms of outdoor thermal comfort in winter and the duration of building sunlight exposure on the winter solstice. Clusters 3, 4, 5, and 6 are more balanced than 1 and 2. To compare the performance of different objectives in each group, please refer to Figure 14.

Table 6. Some schemes of each cluster.

Clustering	Scheme Model
Cluster1
	W_U: −1.868	W_S: 3.971	W_U: −2.628	W_S: 4.118	W_U: −1.227	W_S: 3.890	W_U: −1.791	W_S: 3.954	W_U: −1.717	W_S: 3.996
	S_U: 29.005	S_R: 14.265	S_U: 28.943	S_R: 14.343	S_U: 29.140	S_R: 14.277	S_U: 29.159	S_R: 14.343	S_U: 29.130	S_R: 14.489
Cluster2
	W_U: −5.374	W_S: 3.201	W_U: −5.333	W_S: 3.130	W_U: −4.984	W_S: 3.260	W_U: −6.129	W_S: 3.593	W_U: −5.166	W_S: 3.463
	S_U: 28.302	S_R: 12.169	S_U: 28.229	S_R: 12.124	S_U: 28.765	S_R: 12.416	S_U: 28.342	S_R: 13.095	S_U: 28.443	S_R: 12.789
Cluster3
	W_U: −3.325	W_S: 3.707	W_U: −3.221	W_S: 3.769	W_U: −3.221	W_S: 3.769	W_U: −3.537	W_S: 3.738	W_U: −3.501	W_S: 3.913
	S_U: 28.749	S_R: 13.566	S_U: 28.767	S_R: 13.565	S_U: 28.767	S_R: 13.565	S_U: 28.730	S_R: 13.613	S_U: 28.742	S_R: 13.835
Cluster4
	W_U: −3.961	W_S: 3.336	W_U: −3.506	W_S: 3.373	W_U: −4.403	W_S: 3.246	W_U: −3.064	W_S: 3.327	W_U: −3.084	W_S: 3.327
	S_U: 28.480	S_R: 12.679	S_U: 28.603	S_R: 12.874	S_U: 28.500	S_R: 12.410	S_U: 28.682	S_R: 12.903	S_U: 28.682	S_R: 12.903
Cluster5
	W_U: −2.390	W_S: 3.565	W_U: −2.745	W_S: 3.705	W_U: −2.012	W_S: 3.435	W_U: −2.739	W_S: 3.381	W_U: −1.929	W_S: 3.609
	S_U: 28.839	S_R: 13.385	S_U: 28.843	S_R: 13.551	S_U: 28.845	S_R: 13.145	S_U: 28.741	S_R: 12.966	S_U: 28.902	S_R: 13.430
Cluster6
	W_U: −4.670	W_S: 3.805	W_U: −5.620	W_S: 3.727	W_U: −4.090	W_S: 3.716	W_U: −3.879	W_S: 3.604	W_U: −3.828	W_S: 3.814
	S_U: 28.595	S_R: 13.294	S_U: 28.670	S_R: 13.374	S_U: 28.859	S_R: 13.324	S_U: 28.819	S_R: 12.989	S_U: 28.652	S_R: 13.602

Note: “W_U”: Winter_UTCI; “W_S”: Winter _Sunhours; “S_U”: Summer_UTCI; “S_R”: Summer_Radiation.

shows some of the schemes in various groups.

5. Conclusions

This study proposes a comprehensive methodology for the automated generation and multi-objective optimization of urban block morphologies, driven by thermal comfort requirements. Morphological parameters—such as building height, density, layout type, and block road width—are treated as independent variables. By integrating parametric modeling with environmental performance simulations, the framework enables simultaneous optimization of both indoor and outdoor environmental objectives. Its high flexibility and operability allow for diverse variable-objective combinations, providing a scientific and efficient tool to support climate-responsive urban design.

The optimization results reveal that winter Universal Thermal Climate Index (UTCI) values range from −6.13 °C to −1.18 °C, while summer values range from 28.19 °C to 29.17 °C. The broader variation in winter UTCI indicates greater potential for improving outdoor thermal comfort in winter compared to summer under the current conditions. Correlation analysis among optimization objectives shows a strong synergistic relationship between winter thermal comfort and building sunshine duration on the winter solstice—both can be simultaneously enhanced through appropriate building layout and height strategies. Similarly, summer thermal comfort is positively correlated with reduced building solar radiation, as minimizing solar exposure helps alleviate thermal loads. However, trade-offs are necessary due to evident conflicts: winter and summer thermal comfort objectives compete with each other, as do the winter sunshine duration and summer comfort and radiation indices. These findings highlight the need for seasonally adaptive design strategies and the prioritization of performance goals based on project-specific requirements.

By analyzing the intrinsic relationships between performance objectives and architectural elements, this study identifies the common morphological characteristics of high-performance design schemes for each objective. For schemes with higher winter UTCI values, building layouts typically follow a “low-south, high-north” configuration, dominated by low-rise and mid-rise structures. These schemes often adopt enclosed planar forms, with a density pattern characterized by higher perimeter density (greater than 0.3) and lower density at the center, while inter-cluster road widths remain relatively narrow (around 10 m). In contrast, schemes achieving lower summer UTCI values generally feature high-rise buildings to maximize shading effects, favor point-type layouts, and maintain relatively low overall density (mostly below 0.2) with minimal variation between clusters—conditions that promote ventilation and heat dissipation. Schemes optimized for extended sunshine duration on the winter solstice exhibit a height pattern that is lower in the southeast and higher in the northwest, frequently employing point-type and enclosed plan forms, with wider roads to reduce mutual shading. Lastly, schemes with reduced summer building solar radiation tend to have lower building heights at the four corners of the block, a concentration of high-rise structures in the core area, predominantly point-type or slab-type layouts, uniformly low building density, and consistently narrow roads.

In addition, the study applies the K-means clustering algorithm to analyze all feasible design schemes, revealing significant performance differences across clusters for each optimization objective. This multi-scheme output model overcomes the limitation of relying on a single optimal solution in traditional design workflows. It enables architects to flexibly select design alternatives based on specific project requirements and the relative weighting of performance objectives. By incorporating environmental performance evaluation at the early design stage, the approach significantly improves the efficiency of both scheme optimization and environmental assessment, addressing a key shortcoming of conventional design processes.

6. Discussion

6.1. Theoretical Contributions

This study achieves three key breakthroughs at the theoretical level, significantly enhancing the scientific rigor and systematization of climate-adaptive urban design:

(1) Quantitative Analysis of Multi-Objective Correlation Mechanisms

Traditional urban design often relies on empirical and qualitative judgments to understand the relationship between form and climate. In contrast, this study adopts a quantitative analytical approach and reveals several key interactions: winter thermal comfort (UTCI) exhibits a strong synergistic relationship with sunshine duration (synergy coefficient: 0.777), while summer thermal comfort is positively correlated with solar radiation (synergy coefficient: 0.947). However, significant trade-offs exist between winter and summer UTCI values (conflict index: −0.975), and between winter solstice sunshine duration and both summer UTCI (−0.871) and solar radiation (−0.973). These findings break through the traditional model of “single-objective empirical trial-and-error” by establishing, for the first time, a coupled model involving 12 morphological parameters and four climatic performance objectives. The model uncovers the non-linear interaction mechanisms within the “form–climate” system, providing verifiable quantitative evidence to support the advancement of climate-adaptive urban design theory.

(2) Data-Driven Transformation of Cross-Seasonal Design Guidelines

Traditional design guidelines—such as the fixed sunlight spacing ratio (D/H = 1.2)—lack adaptability across seasons. Through Pareto front analysis, this study extracts season-specific morphological strategies. In winter, a “low-south, high-north enclosed” layout (building density: 0.33; road width: 10 m) improves the UTCI from an average of −8.1 °C to −1.45 °C—an increase of 6.65 °C. In summer, a “high-rise, point-block, low-density” configuration (density: 0.16; height: 93.7 m) reduces UTCI from 32 °C to 28.23 °C—a decrease of 3.77 °C. These data-driven design principles break away from the conventional “one-size-fits-all” approach by establishing precise mappings between climate performance objectives and morphological parameters. This represents a shift in urban design theory from vague, experience-based decision-making to scientifically grounded quantitative methods.

(3) Refinement of Theoretical Frameworks for Temperate Monsoon Climate Zones

Most existing theories focus on extreme climate regions, while this study addresses the distinct four-season characteristics of temperate monsoon climate zones, which cover approximately 22.2% of China’s land area. A dual-season optimization framework—targeting both winter and summer performance—was developed to guide urban design in these regions. By applying K-means clustering (K = 6) to classify 277 Pareto-optimal solutions, the study identified six representative morphological schemes. Based on these results, a novel “dynamic balance” design paradigm is proposed for the first time, bridging the theoretical gap in climate adaptability for temperate monsoon zones and expanding the climatic applicability of urban design theory.

6.2. Practical Implications

This study provides a practical and operable technical pathway for urban planning and design, offering several significant advantages over traditional approaches:

(1) Efficient Multi-Objective Design Workflow

Traditional design processes often require weeks of iterative adjustments to balance conflicting objectives such as solar access and ventilation. In contrast, this study establishes an automated workflow integrating parametric modeling, multi-physics simulation, and genetic algorithm optimization (Figure 5). The optimization cycle for a single design scheme is reduced from several weeks to approximately 350 h across 50 generations, producing a Pareto front with 277 valid solutions that encompass both winter and summer performance. Designers can compare and select optimal schemes in real time using multi-axis indicator distribution charts (Figure 8), greatly enhancing design efficiency.

(2) Precise Generation of Climate-Sensitive Morphologies

While conventional design relies on subjective judgment, this study identifies key morphological parameter thresholds through quantitative analysis. For optimal winter thermal comfort, the design should maintain a building density of 0.3–0.38 and an enclosure degree of ≥75%. For summer ventilation optimization, point-block buildings should account for ≥84% of the layout, with overall density kept below 0.2.

(3) Innovative Application of Dynamic Adaptive Strategies

Static design approaches struggle to accommodate seasonal variations. This study reveals a strong trade-off between seasonal objectives—each 1 °C increase in winter UTCI leads to an approximate 0.975 °C rise in summer UTCI. These findings encourage a shift from fixed-form solutions to dynamic adjustment strategies. Future applications may combine high-rise point-block forms with movable shading devices and phase-change materials to mitigate seasonal UTCI fluctuations and achieve balanced year-round performance.

6.3. Limitations and Future Research

Despite the valuable outcomes achieved in this study, several limitations remain. In terms of model construction, certain physical processes were simplified—for example, the complex interactions of longwave radiation and turbulent airflow patterns within urban blocks were not fully accounted for, which may introduce deviations between simulation results and real-world conditions. Moreover, the study relies primarily on simulation data and lacks validation through field measurements. Future research should incorporate empirical monitoring data to calibrate and refine simulation models, thereby improving result accuracy. The scope of the evaluation could also be expanded to include additional environmental factors such as air quality and acoustic comfort, aiming to develop a more comprehensive urban environmental performance assessment framework. Furthermore, the integration of machine learning techniques, such as deep learning and reinforcement learning, holds potential for enhancing real-time optimization capabilities, increasing both the intelligence and responsiveness of the proposed methodology.

In conclusion, this study lays a solid foundation for sustainable and climate-adaptive urban design. Its methodological innovations and key findings offer significant theoretical and practical value for advancing the fields of urban planning and architectural design.