Optimizing daylight in west-facing facades for LEED V4.1 compliance using metaheuristic approach Authors: Pham Vu Hong Son, Vo Thi Bich Huyen Published: 2023 ABSTRACT This study introduces an optimized design approach for west-facing room façades to improve daylighting while adhering to LEED v4.1 sustainability criteria. Employing parametric modeling, metaheuristic optimization, and validated daylight simulations, the research highlights the African Vulture Optimization Algorithm's success in achieving 100% LEED compliance and superior performance over random models in daylight sufficiency and glare reduction. Light-colored materials and transparent glazing emerged as beneficial for LEED points. Despite computational limitations and the need for empirical validation, this method offers architects versatile and sustainable design solutions. Comparative analysis reveals the algorithm's strong performance, although opportunities exist for refinement. Future research directions include contrasting this algorithm with other optimization methods, focusing on empirical backing, assessing environmental and human-centric impacts, adapting to varied building types and conditions, and examining diverse geographical and material factors. This work advances daylight-integrated façade design, suggesting a more comprehensive framework for building performance optimization. Introduction In recent times, green building design has become increasingly prominent in architectural practices worldwide1. Rating systems such as Leadership in Energy and Environmental Design (LEED) offer benchmarks for sustainable building standards2. A critical aspect of this sustainability focus is the optimization of daylighting, which can both increase visual comfort and decrease electricity consumption3. However, reconciling the advantages of daylight with potential issues such as glare and overheating presents complexities, especially for west-facing facades exposed to western sunlight. Building facades play a significant role in shaping a structure's energy and daylighting characteristics, influencing energy consumption through the modulation of heat, light, and airflow4. Effective daylighting can decrease the reliance on artificial lighting, promote even light distribution, and enrich views, which can contribute to occupant well-being. Metrics like sDA (Spatial Daylight Autonomy) and ASE (Annual Sunlight Exposure) serve as standard tools in the industry to measure daylight performance5. The ongoing objective is to optimize daylight to enhance energy efficiency and comfort while adeptly controlling glare to ensure visual comfort. While numerous studies have addressed facade optimization in various contexts, there remains a specific gap in developing an optimization approach tailored for west-facing facades that aligns with the stringent requirements of the LEED v4.1 daylight criteria. In Asia, pattern ventilation tiles, a culturally significant architectural feature, have re-emerged as a popular design trend, as illustrated by some examples in Fig. 1. This paper aims to contribute to this tradition by proposing an optimization approach for west-facing facades to enhance daylight quality while aligning with LEED v4.1 daylight criteria6. The objective of this work is not to identify an optimal solution, but rather to investigate an approach that can generate varied and advantageous design configurations, improving visual appeal while sustaining compliance with LEED criteria. Furthermore, the proposed methodology is intended for practical application, consisting of a modular and parametric model utilizing commercially available materials. The techniques harness existing algorithms and software platforms to reconcile the advantages of daylighting with the complications it poses.Figure 1Examples of buildings that utilize ventilation bricks. Examples of buildings that utilize ventilation bricks. Literature review Daylighting metrics Prior to reviewing previous relevant research, it is crucial to elucidate the daylighting metrics that will guide this investigation. The Leadership in Energy and Environmental Design (LEED) green building certification system was selected for this study owing to its broad recognition as an industry benchmark2. More specifically, version 4.1 of LEED was employed, which emphasizes passive design and daylighting approaches. The integration of natural daylight provides several benefits, including:Reduced energy consumption through lessened reliance on artificial lighting3. By harnessing natural light sources, buildings can decrease electricity requirements.Enhanced occupant experience through exposure to natural illumination7. Access to daylight has been correlated with improved wellbeing and satisfaction among building inhabitants.Heightened alertness and productivity7. Natural light exposure assists in regulating circadian rhythms and cognitive performance.Potentially lowered absenteeism resulting from these factors. By generating more inviting and engaging spaces with daylight, employee attendance and retention may also increase. Reduced energy consumption through lessened reliance on artificial lighting3. By harnessing natural light sources, buildings can decrease electricity requirements. Reduced energy consumption through lessened reliance on artificial lighting3. By harnessing natural light sources, buildings can decrease electricity requirements. Enhanced occupant experience through exposure to natural illumination7. Access to daylight has been correlated with improved wellbeing and satisfaction among building inhabitants. Enhanced occupant experience through exposure to natural illumination7. Access to daylight has been correlated with improved wellbeing and satisfaction among building inhabitants. Heightened alertness and productivity7. Natural light exposure assists in regulating circadian rhythms and cognitive performance. Heightened alertness and productivity7. Natural light exposure assists in regulating circadian rhythms and cognitive performance. Potentially lowered absenteeism resulting from these factors. By generating more inviting and engaging spaces with daylight, employee attendance and retention may also increase. Potentially lowered absenteeism resulting from these factors. By generating more inviting and engaging spaces with daylight, employee attendance and retention may also increase. The metrics of sDA and ASE were selected for this study as they align with LEED v4.1 daylighting criteria and have become industry standards for assessing daylight performance. sDA indicates the percentage of a space that has sufficient daylight levels for occupants, while ASE helps identify potential glare issues from excessive direct sunlight5. Although other metrics like daylight factor, daylight autonomy, and useful daylight illuminance were considered initially, sDA and ASE were ultimately chosen due to their stronger alignment with LEED requirements. In particular, LEED certified projects must achieve benchmarks:Spatial Daylight Autonomy (sDA) The percent of floor area that meets a minimum daylight illuminance level (300 lx) for at least 50% of occupied hours annually. To earn LEED points, projects must achieve sDA300/50% for at least 40% of the occupied floor area (1 point), 55% (2 points), or 75% (3 points). This study targeted the highest level of 75% sDA6.Annual Sunlight Exposure (ASE) The percent of floor area that receives at least 1000 lx for at least 250 occupied hours per year. LEED v4.1 stipulates an ASE1000/250 upper limit of no more than 10% of the occupied space6. Spatial Daylight Autonomy (sDA) The percent of floor area that meets a minimum daylight illuminance level (300 lx) for at least 50% of occupied hours annually. To earn LEED points, projects must achieve sDA300/50% for at least 40% of the occupied floor area (1 point), 55% (2 points), or 75% (3 points). This study targeted the highest level of 75% sDA6. Spatial Daylight Autonomy (sDA) The percent of floor area that meets a minimum daylight illuminance level (300 lx) for at least 50% of occupied hours annually. To earn LEED points, projects must achieve sDA300/50% for at least 40% of the occupied floor area (1 point), 55% (2 points), or 75% (3 points). This study targeted the highest level of 75% sDA6. Annual Sunlight Exposure (ASE) The percent of floor area that receives at least 1000 lx for at least 250 occupied hours per year. LEED v4.1 stipulates an ASE1000/250 upper limit of no more than 10% of the occupied space6. Annual Sunlight Exposure (ASE) The percent of floor area that receives at least 1000 lx for at least 250 occupied hours per year. LEED v4.1 stipulates an ASE1000/250 upper limit of no more than 10% of the occupied space6. Previous studies Numerous studies have delved into the topic of facade optimization. ElBatran et al.8 investigated the design of an optimal double skin south-facing façade for Egypt's hot climate to align with LEED requirements. However, their research did not encompass the influence of façade geometry, reflectivity, and color on daylight performance. Similarly, Zhang and Ji9 employed parametric tools, such as Rhino and Grasshopper, to optimize apartment window designs using genetic algorithms, yet LEED sustainability metrics were not a focal point of their study. Glassman and Reinhart10 explored the impact of future climate scenarios on facade optimization using parametric design and simulation tools. Their study, while insightful, focused on energy and carbon emissions as the main criteria, neglecting the human-centric aspects of facade performance. Daylight quality, visual comfort, and occupant satisfaction were not considered in their analysis. Shan and Junghans11 developed an evolutionary algorithm called adaptive radiation to optimize building facade design for different climates. Their approach modifies a simple genetic algorithm to achieve near globally optimal solutions with substantially reduced computational cost. Although their methodology efficiently optimizes facades for energy performance related to cooling, heating, and artificial lighting requirements, it does not account for human-centric factors that impact facade design. Bakmohammadi and Noorzai12 proposed a framework to assist architects in evaluating classroom designs, with an emphasis on thermal comfort and sustainability. In a different approach, Hosseini et al.13 highlighted the role of facades in enhancing daylight performance in architectural designs. Their model, while intricate, was tailored for a specific building type, potentially limiting its broader applicability. LEED metrics were not a component of their analysis. Wang et al.14 focused on refining a composite shading system for classrooms in Nanchang, China, using 3D parametric modeling and the NSGA-II algorithm. However, their study did not incorporate sustainability metrics. Tabadkani et al.15 assessed visual comfort by contrasting dynamic facade systems with static ones, targeting the improvement of daylight performance. In another study, Fang and Cho16 delved into facade optimization across different climates, analyzing various design variables for a small office building in cities like Miami, Atlanta, and Chicago. Do and Chan17 offer valuable insights into daylight and glare control for multi-sectional facades. Our research aims to complement their work by directly addressing LEED v4.1 daylight criteria, with a focus on visual comfort and daylight sufficiency. While Do and Chan's study provides a strong foundation, the alignment with LEED standards in our work allows for more direct applicability to LEED-targeted projects. Additionally, by optimizing west-facing facades with ventilation bricks, an architectural feature common in Asia, our study explores a practical application while adhering to sustainability principles. This study may contribute meaningfully to the field of sustainable building design through its blended approach, aligning established LEED metrics with regionally relevant designs. Kizilörenli and Tokuç18 developed a novel responsive façade system targeting daylight optimization for west-facing façades. Their parametrically modeled system focused on maximizing Spatial Daylight Autonomy (sDA) while keeping Annual Sunlight Exposure (ASE) under 10% to reduce visual disturbance. Employing tools like Rhinoceros 3D, Grasshopper, Climate Studio, and Octopus, they effectively improved daylight data from initial models, achieving better user comfort levels. Their approach, although parallel to ours in optimizing daylight control, diverges in several aspects. Notably, their method did not incorporate LEED standards, lacked material variation, and compared outcomes only to a base scenario with no façade, rather than exploring a range of real-life façade designs. This approach, while yielding a range of solutions, omits a direct comparison with commonly used façade designs, potentially limiting the broader applicability of their findings in real-world scenarios. Prior studies have provided important insights into facade optimization, yet some areas remain less explored. Many studies did not examine the integration of facade geometry, reflectivity, and color for daylight performance and occupant satisfaction. Others emphasized energy use and thermal comfort without incorporating LEED metrics. Some works focused on specific buildings or contexts. Building on these contributions, the current study aims to optimize west-facing facades for LEED v4.1 daylight criteria. We evaluate visual comfort, daylight sufficiency, and ventilation bricks—an architectural feature common in Asia. By aligning with LEED standards and focusing on regionally relevant elements, we hope to connect theory and practice, adding to knowledge on sustainable facade design. Nature-inspired metaheuristics: exploring AVOA's potential and performance Metaheuristics have garnered significant attention in the realm of computational intelligence for their ability to offer solutions to complex optimization problems that traditional algorithms struggle with19. These algorithms, often inspired by natural or sociological phenomena, aim to strike a balance between exploration (searching new areas in the solution space) and exploitation (refining current solutions). The ingenuity of these techniques lies in their emulation of specific behaviors observed in nature or society, effectively leveraging the inherent wisdom of these systems. In line with this growing interest in bio-inspired optimization techniques, researchers have recently begun applying such algorithms to real-world optimization problems across various domains. For instance, Pham, Nguyen Dang, et al.20 introduced an enhanced SCA that integrates roulette wheel selection with OBL, demonstrating superior performance over traditional optimization algorithms in various engineering optimization contexts. Son and Nguyen Dang21 presented the MVO model as an effective tool for addressing time–cost optimization issues in construction project management, surpassing other techniques in small-scale applications. Pham et al.22 articulated a proficient scheduling optimization technique for ready-mix concrete (RMC) truck dispatches. At its core is a novel hybrid swarm intelligence algorithm fusing the grey wolf optimizer (GWO) with the dragonfly algorithm (DA). The resultant algorithm excels in performance compared to its standalone counterparts and heralds a leap in multi-independent batch plant cooperation for refined RMC deliveries in construction sectors. Son and Nguyen Dang23 showcased a composite model named the hybrid multi-verse optimizer (hDMVO), which integrates MVO and the sine cosine algorithm (SCA). This model is adept at managing discrete time–cost trade-off (TCTO) quandaries in construction project orchestration. Its prowess shines through benchmark evaluations and its knack for devising superior solutions in large-scale TCTO scenarios for intricate projects. For repeating tasks with several concurrent instances, the adaptive selection slime mould algorithm (ASSMA) is presented by Son and Khoi24. Son and Khoi also provides the mutation-crossover slime mold algorithm (MCSMA) for balancing time, cost, quality, and work continuity in a specific building project25. In addition, there has been extensive research in the field of artificial intelligence. Son and Nam 26 employed Hybrid Sine Cosine Optimization to solve the Transport Project. Hybrid multi-verse optimizer model for a sizable discrete time–cost trade-off problem, by PVH Son and NDN Trinh23. To address the drawbacks of the GWO algorithm, Son and Trang22 introduces HDGM, an unique hybrid optimization model combining the dragonfly algorithm and the grey wolf optimizer. P Vu-Hong-Son et al.27 proposed project schedule optimization according to limited resources using dependency structure matrix and whale optimization algorithm. Son and Lien28 seeks to advance a Blockchain-based approach for meeting small- and medium-sized project delay resolution demands in a prompt and transparent manner. In addition to the array of bio-inspired algorithms already discussed, researchers have also begun examining the foraging behaviors of vultures as inspiration for optimization techniques. The African Vulture Optimization Algorithm (AVOA)29, draws inspiration from vultures' adaptive foraging behaviors. AVOA replicates vultures' success in foraging by considering hunger levels, physical capabilities, flight adaptability, and inter-vulture dynamics. By synthesizing these factors, AVOA offers a robust optimization technique for complex problem landscapes, mirroring vultures' adeptness in searching for nourishment under varying conditions. Key features of AVOA encompass modeling hunger as solution quality, physical strength as exploration radius, adaptive flight through adjustable parameters, and conflict avoidance via separation distance. We have considered the AVOA as our main candidate algorithm for this study, aiming to maximize visual comfort metrics, for the following reasons:Recent development: AVOA, introduced in 202129, is a newly developed algorithm undergoing active refinement30–35, offering opportunities for further improvement and application expansion.Documented efficiency: Preliminary studies have demonstrated AVOA's consistently impressive performance, often surpassing established metaheuristic algorithms on benchmark optimization tasks29,35–37.Versatility: The algorithm exhibits versatility in addressing a wide range of optimization challenges, including those with multiple constraints and complex nonlinear problems35–37. Recent development: AVOA, introduced in 202129, is a newly developed algorithm undergoing active refinement30–35, offering opportunities for further improvement and application expansion. Recent development: AVOA, introduced in 202129, is a newly developed algorithm undergoing active refinement30–35, offering opportunities for further improvement and application expansion. Documented efficiency: Preliminary studies have demonstrated AVOA's consistently impressive performance, often surpassing established metaheuristic algorithms on benchmark optimization tasks29,35–37. Documented efficiency: Preliminary studies have demonstrated AVOA's consistently impressive performance, often surpassing established metaheuristic algorithms on benchmark optimization tasks29,35–37. Versatility: The algorithm exhibits versatility in addressing a wide range of optimization challenges, including those with multiple constraints and complex nonlinear problems35–37. Versatility: The algorithm exhibits versatility in addressing a wide range of optimization challenges, including those with multiple constraints and complex nonlinear problems35–37. In exploring the potential and performance of the AVOA for optimizing façade designs to enhance daylight performance, it is acknowledged that the reasons for choosing AVOA, as outlined earlier, carry a degree of subjectivity. To assess the validity of these reasons and to gain a broader perspective on AVOA's performance, this study will also conduct a preliminary comparison with two well-established metaheuristic algorithms: the Genetic Algorithm (GA) and the Grey Wolf Optimizer (GWO). This comparative analysis is not intended to be an exhaustive or definitive evaluation of these algorithms. Instead, it aims to provide an initial insight into their relative efficiency in the context of facade optimization for daylight sufficiency and LEED compliance, particularly under the constraints of our limited hardware resources. By acknowledging these limitations, this comparison seeks to offer an insight into the performance of these algorithms in our specific setting, rather than delivering a comprehensive and fully validated assessment. Thus, the primary objective is to explore whether AVOA's selection based on the earlier detailed criteria holds up when compared with GA and GWO in a constrained computational environment. The Genetic Algorithm (GA), a popular search heuristic that simulates the process of natural selection, was first proposed by Holland38. However, the groundwork for GA can be traced back to 1975, marking its deeper roots in optimization theory39. This algorithm is widely used for solving complex optimization problems and evolves solutions over generations40. Its adaptability and robustness make GA an appropriate benchmark for comparison, demonstrating its long-standing relevance and efficacy in various optimization contexts. The Grey Wolf Optimizer (GWO), inspired by the social hierarchy and hunting strategy of grey wolves, is a powerful metaheuristic algorithm introduced more recently in 201441. Its ability to explore and exploit the solution space effectively has made it a popular choice in various optimization contexts. This popularity can be attributed to its simplicity, ease of use, flexibility, scalability, and its special ability to balance exploration and exploitation during the search process42. As a more modern benchmark, GWO serves alongside the classic GA for comparison with the AVOA, providing a contemporary perspective in evaluating the efficacy of new optimization methods. In addition to offering preliminary insights into the performance of the AVOA, the GA, and the GWO in optimizing daylight sufficiency in façade designs, this study also seeks to explore the nature of design solutions proposed by each algorithm. A critical aspect of this exploration is understanding the diversity of the solution space each algorithm navigates and the specific preferences or tendencies they exhibit in selecting materials and designs. Methodology To develop an effective optimization approach for west-facing facades, our methodology integrates parametric modeling, a metaheuristic optimization algorithm, and high-fidelity daylight simulations. This allows the exploration of diverse design configurations to identify solutions that enhance daylight performance aligned with LEED v4.1 sustainability standards. The parametric model enables flexibility through key variables like facade geometry, materials, and glazing properties. We employ the African Vulture Optimization metaheuristic algorithm for its strong empirical performance in complex optimization tasks. Validated ClimateStudio simulations provide accurate analysis of daylight metrics for proposed configurations. By strategically combining these capabilities, the methodology aims to generate optimized west-facing facade designs that maximize indoor daylight quality and visual comfort. The goal is an adaptable optimization technique that offers architects expanded design possibilities while ensuring adherence to green building best practices. The following sections delve deeper into the algorithm, software tools, simulation parameters, and metrics underpinning this approach. African Vultures Optimization Algorithm (AVOA) In the following section, we provide a concise overview of AVOA as originally developed and presented by Abdollahzadeh et al.29, including the key formulas and equations from their research. This summary aims to succinctly elucidate the algorithm's fundamental components and operational mechanisms for the reader. Figure 2 illustrates AVOA's process flow in a visual flowchart, spanning from initialization to termination criteria.Figure 2Overview of African Vulture Optimization Algorithm. Overview of African Vulture Optimization Algorithm. Initialization Population initialization: Begin by determining the number of vultures (N) relevant to the problem at hand. Create an initial population consisting of N individuals. Step one: Determination of optimal vulture selection within a group After initializing the population, the fitness of all solutions is calculated. The solution with the highest fitness is the optimal vulture in the first group, and the second highest fitness solution is the optimal vulture in the second group. The remaining solutions are directed towards the best solutions in each group using Eq. (1). This process repeats each iteration by recomputing the entire population's fitness.1\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\mathrm{R}\left(\mathrm{i}\right)=\left\{\begin{array}{c}{BestVulture}_{1}if {P}_{i} =L1\\ {BestVulture}_{2}if {P}_{i} =L2\end{array}\right.$$\end{document}Ri=BestVulture1ifPi=L1BestVulture2ifPi=L2 Equation (1) determines the probability of selecting the designated vultures to direct other vultures towards one of the best solutions in each group. Parameters L1 and L2, measured before the search, have values between 0 and 1 that sum to 1. Equation (2) establishes the selection probability of the best solution in each group using the Roulette wheel method.2\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${P}_{i}=\frac{{\mathrm{F}}_{i}}{\sum_{i=1}^{n}{F}_{i}}$$\end{document}Pi=Fi∑i=1nFi Step two: The rate of vulture starvation Satiated vultures fly farther; hungry ones struggle and get aggressive. Equation (4) models this behavior, signaling when to shift from exploration to exploitation based on satiety rates. In these equations, F is the vulture's satiation, iterationi is the current cycle, and maxiterations is the total cycles. z and h are random numbers affecting the state of the vulture. Optimization problems do not guarantee precise global solutions, risking convergence to local optima. Equation (3) mitigates this by enhancing adaptability. Initial iterations of the AVOA focus on exploration, while later ones execute exploitation. The goal is to tweak Eq. (3) to oscillate between these phases, enhancing AVOA's optimization efficiency. In Eq. (3) w is a fixed parameter affecting the transition between phases. Vultures explore less as iterations proceed. When |F| > 1, AVOA enters the exploration phase; if |F| < 1, it switches to exploitation.3\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$t=\mathrm{h}\times \left({\mathrm{sin}}^{w}\left(\frac{\pi }{2}\times \frac{{iteration}_{i}}{maxiterations}\right)+\mathrm{cos}\left(\frac{\pi }{2}\times \frac{{iteration}_{i}}{maxiterations}\right)-1\right)$$\end{document}t=h×sinwπ2×iterationimaxiterations+cosπ2×iterationimaxiterations-14\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$F=\left(2\times {rand}_{1}\right)\times z\times \left(1-\frac{{iteration}_{i}}{maxiterations}\right)+t$$\end{document}F=2×rand1×z×1-iterationimaxiterations+t Step three: Exploration In the AVOA framework, vultures use two strategies determined by a parameter, P1, ranging from 0 to 1. A random value, randP1 (0–1), selects the strategy: Eq. (6) if randP1 ≤ P1, else Eq. (8) (as in Eq. 5).5\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\mathrm{P}\left(\mathrm{i}+1\right)=\left\{\begin{array}{c}Equation\left(6\right)if{P}_{i} \ge {rand}_{P1}\\ Equation\left(8\right)if{P}_{i} <{rand}_{P1}\end{array}\right.$$\end{document}Pi+1=Equation6ifPi≥randP1Equation8ifPi