ArticlePublisher preview available

Integrating machine learning in digital architecture: enhancing sustainable design and energy efficiency in urban environments

Authors:
Main Abushaikha at Amman Arab University
Main Abushaikha
Mutasem A. Al-Karablieh
Mutasem A. Al-Karablieh
Mutasem A. Al-Karablieh
  • This person is not on ResearchGate, or hasn't claimed this research yet.
Akram M. Musa
Akram M. Musa
Akram M. Musa
  • This person is not on ResearchGate, or hasn't claimed this research yet.
Maryam Almashayikh at Amman Arab University
Maryam Almashayikh
Show all 5 authorsHide
Read publisher preview
Download citation
To read the full-text of this research, you can request a copy directly from the authors.

Abstract and Figures

The following work applies metaheuristic optimization algorithms—PSO, ACO, Genetic Algorithm, and Enhanced Colliding Bodies Optimization (ECBO)—to the optimum design of a sustainable building with respect to prominent metrics such as energy savings, improvement in indoor comfort, and reduction in carbon footprint. These algorithms are applied to a wide dataset that includes variable intensity factors such as window-to-wall variation ratio, HVAC efficiency, and integration of renewable energy. Results also proved that PSO is the fittest strategy to balance energy efficiency and sustainability, with the highest energy savings of 24.1%. Besides, PSO wasn’t just the fastest convergence rate; it also obtained a Platinum LEED certification. ACO was second in order of magnitude, with high energy savings and carbon footprint reduction values, and also obtained the Platinum LEED certificate. The results obtained for GA were positive from the occupant comfort point of view but were slower in terms of energy savings and convergence speed. In contrast, ECBO had the slowest convergence and lowest energy savings, demonstrating the limitation of the application of ECBO for large-scale multi-objective optimization. These results imply that PSO and ACO would be suitable for practical applications linked to urban sustainable design, while GA and ECBO are more suited for niche applications. The obtained results can provide useful guidelines in developing more energy-efficient and sustainable designs for architects, urban planners, and policymakers.
Figures - available from: Asian Journal of Civil Engineering
This content is subject to copyright. Terms and conditions apply.
A preview of this full-text is provided by Springer Nature.
Learn more
Content available from Asian Journal of Civil Engineering
This content is subject to copyright. Terms and conditions apply.
Vol.:(0123456789)
Asian Journal of Civil Engineering (2025) 26:813–827
https://doi.org/10.1007/s42107-024-01224-4
RESEARCH
Integrating machine learning indigital architecture: enhancing
sustainable design andenergy efficiency inurban environments
Ma’inF.Abu‑Shaikha1· MutasemA.Al‑Karablieh2· AkramM.Musa3· MaryamI.Almashayikh1· RazanY.Al‑Abed4
Received: 15 October 2024 / Accepted: 1 November 2024 / Published online: 21 November 2024
© The Author(s), under exclusive licence to Springer Nature Switzerland AG 2024
Abstract
The following work applies metaheuristic optimization algorithms—PSO, ACO, Genetic Algorithm, and Enhanced Colliding
Bodies Optimization (ECBO)—to the optimum design of a sustainable building with respect to prominent metrics such as
energy savings, improvement in indoor comfort, and reduction in carbon footprint. These algorithms are applied to a wide
dataset that includes variable intensity factors such as window-to-wall variation ratio, HVAC efficiency, and integration of
renewable energy. Results also proved that PSO is the fittest strategy to balance energy efficiency and sustainability, with the
highest energy savings of 24.1%. Besides, PSO wasn’t just the fastest convergence rate; it also obtained a Platinum LEED
certification. ACO was second in order of magnitude, with high energy savings and carbon footprint reduction values, and
also obtained the Platinum LEED certificate. The results obtained for GA were positive from the occupant comfort point of
view but were slower in terms of energy savings and convergence speed. In contrast, ECBO had the slowest convergence and
lowest energy savings, demonstrating the limitation of the application of ECBO for large-scale multi-objective optimization.
These results imply that PSO and ACO would be suitable for practical applications linked to urban sustainable design, while
GA and ECBO are more suited for niche applications. The obtained results can provide useful guidelines in developing more
energy-efficient and sustainable designs for architects, urban planners, and policymakers.
Keywords Metaheuristic algorithms· Sustainable building design· Energy efficiency· Carbon footprint reduction·
Occupant comfort optimization
Introduction
This proposed mechanism of embedding machine learning
(ML) into digital architecture gives a chance to transform
sustainable design, energy conservation, housing, and urban
planning. It becomes more imperative each day to design
building and plan cities that will be sustainable and energy
efficient, as urbanization occurs at an ever-quickening pace.
corporating novel, data-based solutions based on machine
learning to overcome these challenges could allow for max-
imizing resource endowments, enhancing energy perfor-
mance in sustainable buildings and more integrated housing
and planning processes to produce future smarter and more
resilient urban environments (Abu-shaikha 2024; Abusaleh
2024; Abid etal., 2024; Hussein etal., 2024)
The use of machine learning for urban and landscape
design has reached a considerable momentum over the past
few years (Raanan etal., 2022) According to Wanger etal.,
(2022), the concept of ML has emerged recently, especially
in regards to sustainability assessment, with literature on
* Ma’in F. Abu-Shaikha
m.abushaikha@aau.edu.jo
Mutasem A. Al-Karablieh
m.karablieh@ju.edu.jo
Akram M. Musa
a.musa@aau.edu.jo
Maryam I. Almashayikh
m.almashayikh@aau.edu.jo
Razan Y. Al-Abed
r.alabed@aau.edu.jo
1 Department ofArchitecture Engineering, College
ofEngineering, Amman Arab University, Amman, Jordan
2 Visual Art Department, School ofArts andDesign, The
University ofJordan, Amman, Jordan
3 Department ofRenewable Energy Engineering, College
ofEngineering, Amman Arab University, Amman, Jordan
4 Department ofArchitecture Engineering, College
ofEngineering, Amman Arab University, Amman, Jordan
Content courtesy of Springer Nature, terms of use apply. Rights reserved.
ResearchGate has not been able to resolve any citations for this publication.
Smart sustainable architecture: leveraging machine learning for adaptive digital design and resource optimization
Article
Full-text available
  • Sep 2024
This study investigates the application of the Fruit Fly Optimization Algorithm (FOA) in enhancing the predictive performance of the Light Gradient Boosting Machine (LightGBM) model for smart sustainable architecture. Key features, including Energy Consumption, Water Usage, Material Cost, CO2 Emissions, and Design Flexibility, were selected using FOA to optimize the model’s predictive accuracy. The FOA-based feature selection significantly improved across all performance metrics: Accuracy increased from 0.85 to 0.88, Precision from 0.80 to 0.84, Recall from 0.78 to 0.82, and the F1-Score from 0.79 to 0.83. Moreover, the Root Mean Square Error (RMSE) decreased from 0.25 to 0.22, while the Area Under the Curve (AUC) improved from 0.76 to 0.8625. These findings underscore the effectiveness of FOA in refining feature selection, thereby enhancing the efficiency and reliability of predictive models in sustainable architectural design. The study highlights the potential of advanced optimization algorithms in developing more adaptive, resource-efficient, and sustainable architectural solutions.
View
Enhancing preservation outcomes for architectural heritage buildings through machine learning-driven future search optimization
Article
Full-text available
  • Jul 2024
Architectural heritage represents an invaluable tapestry of human civilization’s artistic, cultural, and historical milestones. As urbanization and societal dynamics evolve, preserving these architectural marvels becomes increasingly imperative. Drawing upon a meticulously curated dataset of UNESCO World Heritage Sites, we introduce a novel machine-learning model that harnesses the capabilities of XGBoost in conjunction with Future Search Optimization (FSO). Our findings reveal that the XGBoost-FSO hybrid accurately identifies and categorizes at-risk heritage sites. From an architectural preservation perspective, the transformative potential of such predictive analytics is immense. They enable timely interventions, ensuring that heritage structures, each with its unique narrative and significance, endure the tests of time. By bridging the realms of technology and architectural heritage, this research underscores the promise of data-driven strategies in championing the cause of global heritage preservation. Through this interdisciplinary approach, we envision a future where technology acts as the vanguard, safeguarding the architectural legacies of yesteryears for generations to come.
View
Forecasting and managing urban futures: machine learning models and optimization of urban expansion
Article
Full-text available
  • Jul 2024
This paper presents a novel method for predicting and controlling urban growth by combining convolutional neural networks (CNN) with Spider Monkey Optimization (SMO) to efficiently use their combined capabilities. This strategy utilizes sophisticated machine learning algorithms to forecast urban development patterns, while optimization technologies enhance these forecasts by including sustainable urban design ideas. Integration plays a crucial role in tackling the difficulties presented by fast urbanization, including issues related to land use, resource administration, and environmental sustainability in the fields of civil and architectural engineering. The study employs large datasets to assess the efficacy of CNN models in detecting urban expansion patterns and enhances these models with optimization scenarios encompassing land use, traffic flow, electricity efficiency, water resources, and waste management, all with the goal of enhancing urban sustainability. The findings demonstrate a significant improvement in the precision of predictions and the ability to maintain positive results, highlighting the crucial role of integrating machine learning and optimization in urban planning methodologies. This method not only provides urban planners and politicians with strong and effective tools for making decisions but also creates opportunities for implementing creative and sustainable solutions for urban growth. The results provide a substantial contribution to the existing urban planning literature by demonstrating the actual use of these technologies in regulating urban expansion. This will stimulate more study in this rapidly evolving subject.
View
Medium office energy consumption optimization using EnergyPlus
Article
Full-text available
  • May 2024
This research paper studies the impacts of HVAC system scheduling and building envelope material properties on the building energy efficiency in the downtown New York City, metropolitan area, that is highly associated with Urban Heat Island. By utilizing EnergyPlus for whole energy building simulation, the research compares two HVAC operational cases: a baseline case with constant temperature setpoints, and an occupancy-based temperature setpoint schedule. The study also investigated the influence of thermal conductivity variations in stucco exterior wall materials, with three cases, the default baseline thermal conductivity and a 15% variance, respectively. Results indicate that occupancy-driven scheduling significantly reduces energy consumption by approximately 16% annually compared to baseline. Conversely, the influence of thermal conductivity on energy consumption is minimal, with only a 0.35% deviation noted in response to a 15% variation in thermal conductivity. This suggests that both optimizing HVAC scheduling and the thermal conductivity of building envelope materials can save building energy consumption while optimizing HVAC scheduling has a relatively stronger impact than optimizing the thermal conductivity with a 15% variance in downtown New York City.
View
Integrating machine learning in architectural engineering sustainable design: a sub-hourly approach to energy and indoor climate management in buildings
Article
Full-text available
  • Apr 2024
This study predicts building energy use and indoor climate using CNNs and LOA. The research gives architects and engineers real-time building performance optimization tools using sub-hourly data to meet escalating energy efficiency and interior environment needs. This research presents a new way for improving prediction model accuracy and efficiency using CNN's spatial and temporal data processing and LOA's systematic feature selection. LOA and CNN identified the most critical building energy consumption and indoor climate variables, creating a predictive model with enhanced R-squared and Mean Absolute Error. According to the main findings, machine learning can recognize and quantify complex energy use and interior environment characteristics. The LOA-CNN combination provided a predictive model with better accuracy and efficiency, highlighting environmental and operational factors, and offering a granular picture of sustainable building management. This research reveals applications that could change architectural engineering beyond theoretical. This study uses predictive analytics to precisely regulate and optimize building systems, enabling the development of more intelligent, more sustainable buildings that dynamically respond to occupant needs and environmental circumstances, boosting sustainability and well-being. This study's CNN-LOA integration indicates advanced machine learning techniques may be used in sustainable building design, ushering in data-driven energy efficiency and indoor environmental quality innovation.
View
Integrating Machine Learning in Urban Pedagogy: Addressing Homelessness in Skid Row
Article
Full-text available
  • Feb 2024
This paper investigates the application of machine learning in urban and architectural education, with a focus on addressing homelessness in Skid Row, Los Angeles. It presents a case study of an urban design studio utilizing data-driven methods to propose transitional housing solutions, emphasizing the importance of design in the context of social justice. The study explores the use of machine learning and digital cartography for a detailed analysis of Skid Row’s dense homeless population, offering students a thorough insight into urban challenges. The research also identifies the complexities involved in integrating these technologies into educational frameworks, including issues with data accuracy, technical hurdles, and ethical considerations. The paper concludes by advocating for an interdisciplinary, data-informed, and socially conscious approach in architectural and urban design education, highlighting its necessity in preparing students to effectively tackle contemporary urban problems.
View
Developing a two-level machine-learning approach for classifying urban form for an East Asian mega-city
Article
Full-text available
  • Sep 2023
Having had the most rapid urbanization in the world since the 1990s, mega-cities in East Asia featured highly compact and atomized modernist architecture. With densely built modernist architecture and relatively free building regulations, it is challenging to trace the actual development of the whole city. Compared to European cities, their overall urban landscapes are much denser, higher, and functionally mixed. In order to achieve a quicker and more accurate identification of urban forms in mega-cities, this study proposed a two-level machine-learning approach. At the building level, we extracted features from topographic maps and building licenses to automatically classify building types. Four state-of-the-art multi-class classification models were compared. At the block level, we used building types as input data and compared two methods for block clustering. In total 61,426 buildings from Taipei were classified and grouped into 10 block types. Different from Western cities, many of the block types in Taipei were mixtures of different types of buildings. This approach is efficient in exploring new urban form types, especially for emerging mega-cities where block types are previously unknown. The result not only sheds light on the features of East Asian urban landscapes but also serves as important basis of type-based strategic plans for contemporary urban issues.
View
View
An approximate randomization-based neural network with dedicated digital architecture for energy-constrained devices
Article
Full-text available
  • Nov 2022
  • NEURAL COMPUT APPL
Variable energy constraints affect the implementations of neural networks on battery-operated embedded systems. This paper describes a learning algorithm for randomization-based neural networks with hard-limit activation functions. The approach adopts a novel cost function that balances accuracy and network complexity during training. From an energy-specific perspective, the new learning strategy allows to adjust, dynamically and in real time, the number of operations during the network’s forward phase. The proposed learning scheme leads to efficient predictors supported by digital architectures. The resulting digital architecture can switch to approximate computing at run time, in compliance with the available energy budget. Experiments on 10 real-world prediction testbeds confirmed the effectiveness of the learning scheme. Additional tests on limited-resource devices supported the implementation efficiency of the overall design approach.
View
View