Next Article in Journal
Exploring a Dynamic Homotopy Technique to Enhance the Convergence of Classical Power Flow Iterative Solvers in Ill-Conditioned Power System Models
Previous Article in Journal
Optical Investigation of Sparks to Improve Ignition Simulation Models in Spark-Ignition Engines
Previous Article in Special Issue
Research on Multi-Objective Optimization Design of University Student Center in China Based on Low Energy Consumption and Thermal Comfort
 
 
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Review

A Review of Data-Driven Methods in Building Retrofit and Performance Optimization: From the Perspective of Carbon Emission Reductions

1
College of Architecture and Urban Planning, Tongji University, Shanghai 200092, China
2
Key Laboratory of Ecology and Energy-Saving Study of Dense Habitat, Ministry of Education, Shanghai 200092, China
*
Author to whom correspondence should be addressed.
Energies 2024, 17(18), 4641; https://doi.org/10.3390/en17184641
Submission received: 15 August 2024 / Revised: 15 September 2024 / Accepted: 16 September 2024 / Published: 17 September 2024
(This article belongs to the Special Issue Optimizing Energy Efficiency and Thermal Comfort in Building)

Abstract

:
In order to reduce the contribution of the building sector to global greenhouse gas emissions and climate change, it is important to improve the building performance through retrofits from the perspective of carbon emission reductions. Data-driven methods are now widely used in building retrofit research. To better apply data-driven techniques in low-carbon building retrofits, a better understanding is needed of the connections and interactions in optimization objectives and parameters, as well as optimization methods and tools. This paper provides a bibliometric analysis of selected 45 studies, summarizes current research hotspots in the field, discusses gaps to be filled, and proposes potential directions for future work. The results show that (1) the building-performance optimization (BPO) process established through physical simulation methods combines the site, retrofit variables, and carbon-related objectives, and the generated datasets are either directly processed using multi-objective optimization (MOO) algorithms or trained as a surrogate model and iteratively optimized using MOO methods. When a sufficient amount of data is available, data-driven methods can be used to develop mathematical models and use MOO methods for performance optimization from the perspective of building carbon emission reductions. (2) The benefits of retrofits are maximized by holistically taking environmental, economic, and social factors into account; from the perspectives of carbon emissions, costs, thermal comfort, and more, widely adopted strategies include improving the thermal performance of building envelopes, regulating HVAC systems, and utilizing renewable energy. (3) The optimization process based on data-driven methods, such as optimization algorithms and machine learning, apply mathematical models and methods for automatic iterative calculations and screen out the optimal solutions with computer assistance with high efficiency while ensuring accuracy. (4) Only 2.2% and 6.7% of the literature focus on the impacts of human behavior and climate change on building retrofits, respectively. In the future, it is necessary to give further consideration to user behaviors and long-term climate change in the retrofit process, in addition to improving the accuracy of optimization models and exploring the generalization and migration capabilities of surrogate models.

1. Introduction

1.1. Background

Climate change and greenhouse gas emissions (GHGEs) have emerged as critical concerns impacting the sustainable development of global human habitations [1]. Global Carbon Emissions (CEs) are increasing at a rate of about 2% annually [2] and have risen by approximately 10 billion tons over the past two decades [3]. The construction industry contributes ~40% to the global CEs [4], making it a primary driver of global warming [5]. Moreover, with current high urbanization rates and buildings’ long lifespans [6], urban construction has entered a phase of stock renewal, necessitating extensive retrofitting of existing buildings now and in the future [7]. Retrofitting buildings represents a practical approach to significantly reducing GHGEs [8] and offering substantial environmental, social, and economic benefits [9], thus plays a pivotal role in advancing long-term sustainable developments [10].
A building retrofit is an intervention to add new materials or elements to existing buildings [11], encompassing methods like upgrading the building envelope, enhancing mechanical systems, and improving operations and management [12]. It serves not only to facelift old buildings but also as a potent means to attain environmental sustainability [13] and building-performance optimization (BPO) [14]. BPO objectives guide the direction of building retrofit strategies [15], and when balancing competing goals such as CEs, energy efficiency, and thermal comfort [16], predictive models are essential for selecting optimal solutions from a range of alternatives [17].
Building-performance prediction models are typically categorized into physical models based on building operational process simulations and mathematical models employing data-driven methods [5]. Physics-based models offer precise and reliable calculations, yet their complexity in input parameters and transient modeling increase computational costs. In contrast, mathematical models provide rapid and efficient performance predictions but require extensive data for model developments and often lack physical interpretations of building-performance parameters [18].
Traditional approaches to retrofitting individual buildings typically rely on architects’ knowledge and experience to enumerate and compare options. However, this often faces challenges due to the vast search space and the difficulties in exploring all potential alternatives exhaustively. Advances in computer technology have introduced data-driven methods to the retrofit process, leveraging computer-assisted sampling and iterative calculations for automated evaluations of retrofitting options to derive optimal solutions [19]. Contrasted with traditional methods, this approach overcomes constraints of time and space [20], thereby enhancing the accuracy and efficiency of decision-making processes [21].
Due to the rising complexity of retrofit standards and objectives, surrogate models (also known as metamodels) have garnered significant attention in recent years for their ability to integrate the strengths of both physics-based and mathematical models. These models are trained using a limited set of simulation data as samples. Following parameter adjustments and validations, they can swiftly generate highly accurate results with appropriate input datasets [22], omitting extensive simulations and substantially cutting down calculation times.
Research on building retrofits focusing on Carbon Emission Reductions (CERs) has extensively utilized BPO based on data-driven methods [23,24,25]. While the optimization procedures in various studies share similarities, differences exist in the methods and tools employed. Thus, it is essential to review existing research, comprehend the development and evolution of the focal points of current research, summarize the characteristics and mechanisms behind the most frequently utilized as well as promising new methods and tools, and provide suggestions and guidance for future research and practice towards the design optimization of CER-based building retrofits.

1.2. Previous Reviews

Previous studies have explored topics related to building retrofits, building CERs, and data-driven methods. Addressing current challenges in building retrofits, scholars have offered theoretical guidance and practical recommendations for implementing retrofit projects, choosing retrofitting tools, and formulating retrofit standards. These efforts encompass perspectives such as sustainable retrofits [14], energy-saving retrofits [26], and decision-making methods [27]. In retrofit projects aimed at CERs, researchers primarily assess the environmental impact across the building’s life cycle [11], emphasizing elements like building envelopes, floor plan layouts, and natural ventilation strategies [28] to minimize both Operational Carbon Emissions (OCEs) and Embodied Carbon Emissions (ECEs). Furthermore, data-driven methods play a crucial role in addressing complex issues such as energy predictions and control within buildings [5], contributing to revealing energy-saving potentials. The surrogate models developed through data-driven methods are applied in Building-Performance Simulations (BPSs) and optimization processes to enhance computational efficiency [22].

1.2.1. A Review of Building Retrofit Research

Jagarajan et al. [13] asserted that green buildings can significantly enhance environmental sustainability across design, construction, and maintenance phases, thereby reducing the environmental influence of existing buildings. Their systematic literature review and analysis of green building retrofit initiatives noted a paucity of research concerning critical factors influencing the implementation of retrofit projects, emphasizing the need to focus on the problems faced by stakeholders, such as the lack of financial incentives from the government and the limited availability of green material and technology. Nielsen et al. [14] evaluated decision support tools applicable to building retrofit projects, discussing their application in establishing sustainable development goals, formulating design strategies, and assessing performance. They advocated the use of sustainable standards to provide clearer criteria for screening existing buildings. In a study focused on China, Liu et al. [26] examined obstacles and challenges in building energy-saving retrofit practices and analyzed the evolving national and local incentive measures. Their work contributed to the theoretical foundation of energy-saving policies and standards, proposing a data-driven decision-making system to guide optimal retrofit measure selections. Pombo et al. [29] conducted a review of residential building retrofit research, highlighting the variability in methods for evaluating energy-efficiency measures. They underscored the importance of adopting a life-cycle approach, including methodologies like life cycle assessments (LCAs) and life-cycle costs (LCCs) to identify optimal solutions and assess the retrofit potential of residential buildings. These studies emphasized the multifaceted challenges and opportunities in green building retrofit initiatives, offered insights into enhancing sustainability practices, and supported decision-making framework developments.

1.2.2. A Review of Building CER Retrofit Research

Vilches et al. [11] conducted a comprehensive review on building retrofits and environmental assessments using LCA methods. Their findings highlight that the construction and operational phases are most frequently analyzed, with a primary focus on impact categories such as the global warming potential (GWP) and embodied energy. Mostafavi et al. [28] reviewed 48 studies from 2005 to 2020 investigating energy and carbon performance in high-rise buildings across diverse climates. They identified the significant potential for energy-saving through improvements in envelope designs, floor plan optimizations, and the use of natural ventilation. This study also emphasized strategies to reduce OCEs and ECEs by enhancing thermal performance and integrating recycled materials into building constructions. Li et al. [30] examined sustainable retrofit practices in subtropical high-density cities like Hong Kong, emphasizing the variability of CER outcomes across different building types and environmental conditions. For high-rise and multi-story residential buildings, this study underscored the importance of targeting energy-efficiency improvements in public areas and optimizing building envelopes to meet CER goals. Aghamolaei et al. [31] conducted a research study on decarbonization strategies for university campuses under various climatic conditions. Their research emphasized the critical role of implementing low-carbon and energy-saving measures across campus infrastructures. This review identified key parameters including spatial planning, landscape integration, renewable energy adoption, building envelope enhancement, and sustainable transportation as pivotal for achieving CER goals. Their study also anticipated the Internet of Things (IoTs) and the integration of data-driven technologies to facilitate campus decarbonization efforts through enhanced planning and design processes. These studies contribute to understanding effective strategies and technologies aiming at reducing CEs in buildings and urban environments, providing insights into tailored approaches for different building types and climatic contexts.

1.2.3. A Review of the Application of Data-Driven Methods in Building-Performance Analyses

Wei et al. [5] conducted a thorough review of data-driven methods applied in building energy analyses, encompassing predictive and classification techniques. Their findings underscore the efficacy of these methods in energy load forecasts and modeling, facilitating comprehensive assessments of macro-level energy-saving potentials aligned with consumer demands and the formulation of sustainable urban development strategies. Sun et al. [32] reviewed multi-step energy forecasts using data-driven models. Their comprehensive analysis encompassed feature engineering, model types, and anticipated outcomes, proposing future research directions in predictive capabilities and energy model controls. Roman et al. [16] systematically reviewed the literature on metamodels based on Artificial Neural Networks (ANNs) and BPS. They detailed the methodology for generating ANN-based metamodels, covering stages from sample construction to model training and validation. The authors suggest that for the reproducibility of research results, developers of the metamodel should try to publicize sample databases, which will facilitate more in-depth analyses and wide applications in the future. Westermann et al. [22] explored the application of metamodels in sustainable building design research. They highlighted statistical models as effective surrogates for detailed simulations, offering computational efficiency and mitigating barriers in BPS. Their study synthesized successful applications of metamodels in architectural designs to guide their practical implementation. These studies contribute to advancing methodologies in energy predictions and sustainable analyses through the application of data-driven techniques and metamodeling approaches.
Existing reviews primarily focus on enhancing building energy efficiency or providing decision-making guidance, with limited analyses from the perspective of CERs in building retrofit approaches. Those that discuss CERs tend to compare retrofit strategies from the perspective of different building types or building LCAs, with less discussion on the application of data-driven methods. Reviews of data-driven methods and metamodel applications emphasize energy predictions and sustainable building designs yet rarely integrate these approaches into a comprehensive framework for BPO.
Therefore, based on a systematic review, this article explores the relevance of CERs, building retrofits, and data-driven methods. It introduces research hotspots, compares various data-driven methods applied to enhancing building performance, discusses potentials and challenges in their application in building CER retrofits, and provides suggestions on the selection of appropriate optimization methods from the perspective of CERs for different retrofit conditions and scenarios. It aims to offer an overview of the BPO process using data-driven methods in CER-based retrofits and provide a timely and valuable reference for future research.

1.3. Outline and Structure of This Review

The rest of this article is structured as follows: Section 2 outlines the literature search and selection process (Section 2.1) and methods deployed for bibliometric analyses (Section 2.2) and summarizes the general process of BPO based on data-driven methods (Section 2.3). Section 3 details the methodology for constructing BPO models, covering the simulation method (Section 3.1) and the machine learning method (Section 3.2) and performance indicators and optimization parameters selected during the optimization process. Section 4 summarizes the optimization methods and tools (Section 4.1), as well as the decision-making methods for post-optimization processing (Section 4.2). Building upon these sections, Section 5 discusses the current status and challenges of building retrofit methods (Section 5.1, Section 5.2 and Section 5.3) and outlines future research directions (Section 5.4). Section 6 provides a summary of the entire article.

2. Literature Screening and Bibliometric Analysis

2.1. Literature Search and Screening

The literature search and screening process followed the guidelines of the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) [33], ensuring a transparent, consistent, and comprehensive systematic literature review (SLR) [34]. This process comprised four phases: identification, screening, eligibility, and inclusion. This study conducted an extensive literature search across prominent databases in natural sciences and engineering research, including Scopus, Web of Science (WoS), and Science Direct (SD) [35]. To maximize the number of retrieved articles, the initial search avoided terms like “data-driven”, which were later applied during literature screening. Keywords encompassed three categories related to “building retrofit”, “optimization”, and “carbon emission”, alongside their synonyms, as outlined in Table 1. Each keyword group was combined using the Boolean operator “AND” during the search process.
Firstly, filtering criteria were applied based on the publication year, document type, and language during the initial search. Secondly, duplicate articles were removed, and the remaining results underwent screening based on title, abstract, and keywords. Articles that aligned with the scope of this paper were then subjected to a full-text review. Finally, all references were meticulously checked, and relevant ones meeting the criteria were included. The screening process adhered to the following criteria:
  • To ensure timeliness, articles published within the last 20 years (2003–2023) were selected. Journal articles, known for their rigorous peer-review process [36], were prioritized for their representation and impact in this field [37]. Conference papers, dissertations, and non-English language publications were excluded, retaining only English-language journal articles. Additionally, to facilitate the focus on research methodologies and processes, review papers were omitted from consideration.
  • The literature reviewed addresses CEs during the operational phase of buildings, covering life-cycle carbon emissions (LCCEs), the GWP, and building environmental effects. Studies focusing solely on energy consumption (EC) without converting it to CE objectives were excluded due to differences in concepts and calculation methodologies. Similarly, the literature exclusively discussing CEs during building construction or other phases was also omitted from consideration.
  • The research must incorporate one or more data-driven methods, such as statistical analyses, optimization algorithms, or machine learning (ML) to optimize building performance. Studies that merely listed and compared retrofit plans without employing these methods were excluded.
  • The literature must address the impacts of envelope retrofit measures on the overall building performance. Studies exclusively focusing on the operations of building mechanical systems, energy structure predictions, and similar topics were excluded. Additionally, articles concentrating solely on specific local building components like curtain wall retrofits and structural seismic performance optimizations were also excluded.
Figure 1 illustrates the literature search and screening process, which was applied to retrieve 611 articles from the database. Following screening and supplementation based on the outlined criteria, 45 articles were selected for a detailed analysis. Based on their content, the relevant studies are summarized in terms of building location and type, optimization objectives and variables, and data-driven methods such as machine learning, sensitivity analyses, optimizations, and decision-making methods, as shown in Table 2, Table 3 and Table 4.

2.2. Literature Statistics and Bibliometric Analysis

We present a statistical analysis on the quantitative trends, research locations (Figure 2), types of buildings studied, and data-driven methods used (Figure 3) across all reviewed articles. Figure 2 illustrates that due to rapid advancements in computing applications and the growing maturity of building retrofit research [27], significant research contributions have emerged over the past decade. Over time, the volume of publications has steadily increased, underscoring the field’s substantial research value and potential for development [80]. Geographically, the majority of studies originate from Europe (n = 21) and Asia (n = 17), with a limited representation of regions such as Africa, South America, and Oceania. Notably, there has been a surge in research output from China since 2022. The location of research and local climate conditions play pivotal roles in building retrofit strategies [81], and the diverse environmental and climatic contexts in Europe and Asia contribute to the breadth of study cases reported in these regions [82].
Figure 3 reveals that residential buildings dominate the literature, reflecting their large scale, high EC, and pollution levels and pressing the need for CER retrofits. Among public buildings, office buildings exhibit more pronounced retrofit effects compared to other commercial buildings due to their stable occupancy intensity and schedule, contrasting with the high population mobility typical of commercial spaces. In addition, each article adopts one or more types of data-driven methods. More than 90% of the reviewed studies use optimization methods for analysis, while the application of machine learning methods is less than 20%, which needs further exploration.
A bibliometric analysis provides insights into the current status and developmental trends within a specific research field, capturing its dynamics and directions [83]. Utilizing scientific meteorological mapping tools allows for a detailed analysis of the literature [84]. Tools such as CiteSpace [85] enable citation burst analyses, visually depicting significant changes and trends in the evolution of a research field [86].
CiteSpace 6.3.R1 was employed to conduct a citation burst analysis (Figure 4). A citation burst analysis highlights trends in disciplinary developments based on the evolution of subject terms [87]. Figure 4 displays the top nine keywords sorted by their initiation year, with the red-line segments indicating periods of significant citation [88]. The early literature extensively discussed “genetic algorithm” (2011–2015), whereas recent attention has been paid to keywords like “embodied carbon” and “machine learning” (2021–2023), indicating a gradual shift towards addressing CEs in building retrofits because of global warming. Moreover, data-driven methods such as “machine learning” began gaining attention in building CER retrofit research around 2021.

2.3. General Process of BPO Based on Data-Driven Methods

Through a literature review and analysis, it was determined that in the CER retrofit, BPO based on data-driven methods typically adheres to a consistent process. While minor variations exist in the specifics of methods across the literature, the fundamental steps remain the same. These steps are the acquisition of building information, the construction of BPO models, and the screening and decision-making of optimization solutions, as illustrated in Figure 5.
Figure 5 outlines the general process as follows: (1) A suitable retrofit building type is selected, appropriate samples are classified and selected, and pertinent data, including the building form and indoor environmental parameters, etc., are collected through investigation. (2) A physical model is constructed using the above data through simulation methods, and the model accuracy is validated through a comparison with actual data, such as EC records. Optimization parameters and objectives are identified, functional relationships are established between parameters and objectives through simulations and calculations, and an optimization model is developed. Based on the computational efficiency of the model, the choice of whether to use data-driven methods, such as machine learning, is made to further transform it into a surrogate model. When the amount of data is sufficient, a mathematical model can be directly generated based on the collected building information data, combined with optimization objectives and parameters. (3) Optimization algorithms and decision-making methods are integrated to select the retrofit strategy and achieve the optimal solution aligned with retrofit requirements.

3. The Construction BPO Models

The BPO process, shown in Figure 5, can be divided into two categories: (1) the establishment of physics-based BPS protocols (or workflows) based on multiple combinations of site and building energy-carbon-related variables (parameters), and the generated datasets can be either directly processed in MOO algorithms or trained and validated as a surrogate model and then processed with MOO and (2), the development of a mathematical model using regression or a machine learning method, with a sufficient amount of available empirical data, and this is processed with MOO for the optimal solution(s) from the perspective of CERs.

3.1. Optimization Model Based on Physical Simulation Methods

3.1.1. BPS Tools

To construct an optimization model using physical simulation methods, simulation tools are needed to compute building-performance metrics, such as EC calculations based on heat transfer or thermodynamic theories [89]. BPS tools are being made to be more user-friendly by equipping the simulation engines with certain kinds of graphical user interfaces (GUIs) [90]. An analysis of the BPS tools used in the reviewed papers (Table 5) indicated that DesignBuilder [91], OpenStudio [92] (a plugin for SketchUp [93]), and Honeybee [94] (a plugin for Grasshopper [95]) are the primary tools used in simulation research. A comparative study evaluated the strengths and weaknesses of widely used simulation tools and recommended employing a combination of tools and comprehensively assessing data needs and model suitability to enhance modeling accuracies [96].

3.1.2. Model Calibration

After the physical model is constructed, it is usually necessary to calibrate the model based on monitored weather data and field-measured energy data [97]. The primary calibration process involves adjusting input parameters like heating and cooling set temperatures, indoor occupancy schedules, etc., guided by user experiences and energy consumption records [98].
The model’s accuracy is typically assessed by comparing errors between simulated and actual ECs, commonly evaluated against mean bias errors (MBEs) and coefficient of variation of root mean square error CV (RMSE) criteria as defined by ASHRAE Guideline 14 [99], IPMVP [100], and FEMP [101] (Table 6). A smaller error indicates a greater similarity between the physical model and the operational conditions of the reference building [102]. It is worth noting that the above calibration criteria do not take into account the uncertainty of input parameters, which can be further explored through methods such as energy auditing or Bayesian optimization [103].

3.2. Surrogate and Mathematical Models

Unlike physical simulation methods, machine learning methods use mathematical analyses to establish a functional relationship between input and output data in a black-box model. In the BPO process, the surrogate model converted from the optimization model is applied to the optimization process for pre-processing steps [104], which can expedite the generation of samples [105] and decrease computational costs while preserving accuracy [106]. By utilizing historical weather data, energy profiles, and existing building information, the mathematical model generates output values for each retrofit objective [23], which can be effectively applied to large-scale bottom-up building retrofit analyses, such as community and urban scales [24]. However, it demands extensive and accurate historical and statistical data as a foundation [107]. Moreover, predictive variables are typically limited to certain thermo-physical or equipment parameters, constraining the ability to assess the potential for energy savings from retrofits or technological upgrades [108].
Machine learning has emerged as a crucial tool in recent years, leveraging artificial intelligence (AI) to tackle complex tasks and processes in the building sector [109]. It learns from vast datasets and a relatively small number of input features [110] and utilizes learned information for predictions [111]. The process of converting optimization models into surrogate models using machine learning methods typically involves four main steps: data collection and sampling, data pre-processing and model selection, model training and hyperparameter optimization, as well as model validation [112]:
  • The samples of input data are selected using sampling methods, and output data are computed through physical model simulations. Typically, the sample size should range from 10 to 100 times the number of input parameters [113], although this can vary depending on the model’s complexity [114].
  • Continuous and discrete variables are distinguished, and both input and output variables are standardized and normalized to ensure a comparability of the data. Typically, the correctly formatted data are divided into training and testing sets in an 80%-to-20% ratio [115], after which an appropriate mathematical model is selected [116].
  • The model is trained, and to prevent overfitting (where the model performs well on the training datasets but struggles with out-of-sample data), hyperparameter optimization is needed to balance variances and biases. When choosing hyperparameters, strategies like a grid search can be employed [117], with cross-validations serving as the scoring method [118].
  • The model is validated, and various metrics are chosen to assess its accuracy. Common evaluation indicators include the mean absolute percent error (MAPE), mean absolute error (MAE), and CV (RMSE) [119], with CV (RMSE) being particularly favored, due to its ability to provide a unitless measurement, which facilitates straightforward comparisons of indicators [120].
In constructing surrogate models, various machine learning methods have been utilized in relevant studies of Table 3, including ANNs [24], gradient boosting regression trees (GBRTs) [61], and multiple adaptive regression splines (MARSs) [77], with ANNs being the predominant approach. An ANN is composed of multiple nodes interconnected in a network [121], including input, hidden, and output layers [122]. This method excels in revealing intricate relationships between input and output data [123], outperforming comparable ML algorithms in processing high-dimensional or nonlinear data, as well as exhibiting a superior prediction accuracy and generalization capabilities [124]. Sharif et al. employed an ANN to develop a surrogate model trained and validated using EC data from DesignBuilder, which was applied thereafter to assess ECs, LCCs, and LCAs in retrofit planning. Their findings demonstrate a strong correlation between surrogate and simulation model outputs, and the computation time can be significantly reduced from 170 h to 150 s [45].
In addition, two articles in the literature chose to explore building retrofitting strategies by establishing mathematical models. Studies [24,61] explored retrofit strategies using 12,806 residential buildings in Zurich and 4900 non-residential buildings in the UK, respectively. The researchers selected a GBRT and an ANN as machine learning methods to directly establish mathematical models, respectively, and the calculation accuracy of the models reached over 90%.

3.3. Optimization Objectives and Parameters

In the process of constructing optimization models, it is necessary to determine optimization objectives and parameters according to the research problem and purpose. The optimization objectives can reflect the building performance before and after the retrofit, while the optimization parameters represent the retrofit means and methods to be taken.

3.3.1. Building-Performance Indicators

The articles reviewed in this paper focus on optimization objectives related to building CEs, encompassing LCCEs and OCEs, among others. Given that OCEs constitute more than 70% of the life cycle [125], some studies calculate CEs specifically during the operational phase rather than across the entire life cycle [30]. However, retrofit measures involve additional materials and components, leading to increased ECEs [126], which must be carefully considered when assessing environmental impacts [29]. Furthermore, certain indicators can gauge the built environment performance and indirectly reflect CEs, such as the GWP, which is used to measure the heat-trapping potential of greenhouse gases in the atmosphere and is expressed as CO2-eq in calculations [127].
The built environment profoundly impacts personnel comfort, productivity, and well-being [128]. Therefore, the process of building retrofits must integrate environmental, economic, and social criteria to maximize sustainable benefits [129]. Research indicates that ECs, costs, and thermal comfort are the most frequently selected objective functions in optimization studies [130]. Social criteria, such as thermal comfort and indoor environmental quality, enhance the quality of life and satisfaction [131]. Figure 6 and Figure 7 illustrate a heatmap distribution and frequency ranking, showing the co-occurrence of building CE-related objectives with other research goals in the relevant literature of Table 2. It is evident that optimizing building CEs typically considers enhancing ECs, reducing retrofit costs (RCs) and operational costs (OCs), and improving indoor thermal comfort at the same time. There is also significant research focusing on optimizing CEs and costs throughout the building life cycle.

3.3.2. Optimization Parameters and Variables

The CER retrofit of buildings necessitates a comprehensive consideration of multiple variables to enhance the overall building performance. The retrofit variables of the literature listed in Table 2 typically fall into two categories: envelope-related variables and other variables (Figure 8). Envelope variables encompass vertical opaque envelopes (interior and exterior walls and doors); horizontal opaque envelopes (ground, floor, ceiling, and roof); as well as transparent envelopes (window-related variables). Others primarily include building forms (such as atria and corridors); sunspaces; sunshade components; building equipment and energy systems (HVAC, lighting, and ventilation systems); and renewable energy systems (like photovoltaic (PV) modules and solar thermal devices). Unlike new building designs, considerations regarding building orientation and form parameters are generally unnecessary in retrofits. Figure 8 illustrates that discussions in the literature often focus on building equipment and energy systems, the window thermal performance, the thickness and type of exterior wall insulations, the thickness of roof insulations, and the area covered by PV panels. Transparent enclosures, although typically smaller in area, exhibit greater heat gains and losses compared to opaque enclosures [132], contributing approximately 20% to 40% of the total [127]. Moreover, rural areas often present greater potential for implementing renewable energy technologies (RETs), facilitated by factors such as accessibility and environmental conditions [133]. Therefore, active and passive measures should be comprehensively utilized during the retrofit to make full use of renewable energy while improving the building performance.
A sensitivity analysis is used to evaluate the correlation between input and output parameters within a model, assessing how changes in inputs influence outputs [134]. It can also be regarded as an optimization search method [135]. For complex scenarios with multiple parameters, a sensitivity analysis can rank the importance of variables [136] and narrow parameter ranges before optimization to simplify model complexity [137]. Before a sensitivity analysis, data sampling is essential to ensure suitable samples, which is often accomplished using Latin hypercube sampling (LHS) [138]. This is a method of stratified random sampling [139] that maintains the representativeness and balance of data while significantly reducing the sample size [140]. Table 7 shows the classification of sensitivity analysis methods and their applications in the literature, with global sensitivity analyses, such as the regression-based standardized regression coefficient (SRC) method [67], the variance-based Sobol method [79], and the screening-based Morris method [70], dominating. The selection of sensitivity analysis methods should consider both model applicability and computational costs comprehensively.

3.3.3. Constraints on Optimization Objectives and Parameters

Retrofitting existing buildings has certain constraints such as geometric shapes, site layouts, functional zones, and structural limitations [141]. These constraints confine the range of feasible variables [73], compelling designers to select retrofit measures within these boundaries. Additionally, constraints often stem from regulatory norms, industry standards, or empirical data. For instance, the total area of solar heater and rooftop PV panels cannot exceed the overall roof area [39]. Design specifications also govern aspects like window-to-wall ratios [55], while budgetary limits constrain retrofit project costs [60]. For CE targets, the ECEs of retrofit measures need to be less than the OCEs that are reduced as a result of an improved building performance. The constraints restrict the optimization range of complex retrofit parameters, prompting the efficient search for solutions that meet requirements.

4. Optimization and Decision-Making Process

4.1. Optimization Process Based on Data-Driven Methods

By reviewing the optimization objectives and parameters, it becomes evident that the process of decarbonizing building retrofits involves numerous variables impacting the building performance. To enhance efficiency, a systematic and effective optimization process is essential [142]. Mathematical methods such as optimization algorithms can be used to sample, screen, and iteratively refine retrofit strategies, so as to efficiently identify optimal solutions. Figure 9 illustrates the uses of various optimization methods in the reviewed articles listed in Table 4, with the non-dominated sorting genetic algorithm II (NSGA-II) being the most frequently employed [143].
The exhaustive search method ensures finding the optimal solution but faces an exponentially growing search space with increasing variables, which is often impractical for many real-world optimization problems. A case-by-case analysis through sample screening still relies on enumerated comparisons after sampling and is less utilized in building optimizations [47]. A study employed statistical methods like a regression analysis to model relationships between objective functions and retrofit variables [144]. However, due to the complexity and nonlinearity of BPO problems, the applicability of statistical models may be limited to certain scenarios [145]. To mitigate computational costs and achieve automated iterations and the selection of the optimal solution, researchers have increasingly turned to Evolutionary Algorithms (EAs) for solving optimization problems.
An EA is a population-based random search algorithm [146] known for its faster computational speed, greater accuracy, and enhanced adaptability compared to direct search methods. Including genetic algorithms (GAs), the ant colony optimization (ACO), and the particle swarm optimization (PSO), an EA is effective in navigating the complex search spaces related to building optimization problems [147].
A GA is effective for solving nonlinear and discontinuous problems in building retrofits [148]. It enables a simultaneous search across multiple points in space, thereby reducing the possibility of converging to the local minimum [149]. NSGA-II stands out as one of the most efficient EAs [150], though its performance diminishes with problems involving four or more objectives [151]. NSGA-III [152] offers notable advantages over NSGA-II by enhancing its capability to handle multiple objectives and achieve superior solution distributions [153]. Son et al. examined the optimization of ECs, CEs, RCs, and thermal discomfort hours (TDHs) in public building retrofits, comparing NSGA-II, PSO, multi-objective evolutionary algorithms based on decomposition (MOEA/D), and NSGA-III algorithms. Their findings indicate that NSGA-III exhibits superior diversity and convergence [54]. Furthermore, research has shown that the strength Pareto evolutionary algorithm2 (SPEA2) and NSGA-II perform similarly in solving MOO problems [154], but SPEA2 tends to outperform NSGA-II in high-dimensional objective spaces involving three to four objectives [155].
Similar to GAs, PSO belongs to population-based evolutionary algorithms with meta-heuristic characteristics. However, it does not employ computational processes such as crossovers and mutations [156]. The principle of PSO involves particles searching for optimal solutions by continuously moving through multidimensional search spaces [157].
An optimization analysis necessitates numerous evaluations to achieve near-optimal solutions, leading to relatively extended processing times. Hence, the performance of optimization algorithms plays a crucial role in the efficiency of the optimization process, prompting the selection of suitable methods tailored to specific research requirements [90].
Comprehensive optimization tools integrate multiple algorithms, harmonizing the optimization process. Table 8 illustrates the application of optimization tools in the literature and reveals that Python and MATLAB are the predominant general programming tools utilized in research.
Python is a high-level interpreted programming language. Its ease of learning and vast library of tools have popularized its use in data science and programming [158]. In the context of BPO, Python serves as an application platform for data-driven methods, making the implementation of optimization algorithms more convenient and feasible. For instance, Shadram et al. [58] utilized the pymoo package in Python to implement the NSGA-II algorithm and focused on optimizing the life-cycle energy consumption (LCE), LCCE, and LCC of a conceptualized apartment in various locations in Sweden through retrofit strategies. This study integrated performance simulation and optimization processes using Python and EnergyPlus extension plugins within Grasshopper, establishing a complete and fast optimization process.
MATLAB [159] is a versatile tool for algorithm development, data visualization, and program interaction. The MATLAB Optimization Toolbox facilitates collaborations with other programs, making it an ideal numerical environment for optimization tasks that rely on third-party simulation programs. However, the programming language demands of MATLAB and the complexity of integrating simulation programs can pose challenges for architects. Specialized optimization platforms, like jEPlus + EA [160], can integrate energy simulation programs effectively. Despite offering fewer optimization algorithms, they excel in combining simulation tools and optimization methods with less programming demands, thus enhancing accessibility for architects.
Figure 10 illustrates the relationship among the simulation tools, optimization algorithms, and optimization tools as utilized in the literature. This figure highlights the broad applicability and interaction of tools and methods such as DesignBuilder, NSGA-II, and Python, facilitating diverse applications in research.

4.2. Solution Set Evaluation and Decision-Making Methods

For MOO problems, due to conflicting objectives constrained by each other, the resultant solutions often form a set rather than a single solution, known as the non-dominated or Pareto optimal solution set. The spatial distribution of this set is referred to as the Pareto front. Figure 11a,b illustrate typical examples of the Pareto front for two and three objectives, respectively [161]. Throughout the optimization process, each generation of solution sets requires an evaluation based on metrics such as convergence, uniformity, and diversity [162], with assessment indicators including the inverted generational distance (IGD) [163] and the hypervolume index (HI) [164], among others. The quality of the solution set improves as it approaches the true Pareto frontier, exhibiting better uniformity and a wider distribution.
Each solution within the Pareto front represents a non-dominated outcome, necessitating post-processing decision-making methods to identify the optimal solution. The technique for order preference by similarity to an ideal solution (TOPSIS) [165] and the Utopia-point method are two commonly employed approaches. TOPSIS ranks solutions by evaluating their distances to both positive and negative ideal solutions, accommodating different weights assigned to objectives based on decision-makers’ preferences to derive optimal solutions. Song et al. used the entropy-based method to assign weights to indicators and calculated the information utility value of each indicator based on information entropy, which avoids the influence of human factors compared with subjective empowerment methods. According to the entropy-based method, the GWP, LCCs, and TDHs were assigned weights of 37.29%, 34.17%, and 28.54%, respectively, and were applied to the TOPSIS calculation. The GWP of the final optimized solution can be reduced by 2720 kg/m2 (about 61%) [55].
The Utopia point serves as an ideal reference within the search space, representing the optimal solutions across all objective functions. Solutions on the Pareto front are assessed by their proximity to this point, with the closest solution indicating the optimal compromise when all objectives are equally weighted. Figure 12a,b depict examples of the TOPSIS [166] and Utopia point [167] methods in 2D coordinates.

5. Discussions

5.1. The Research Status Quo

The increasing prevalence of relevant research in recent years may be attributed to a deeper understanding and increased interest in the significance of CERs within the building construction industry. Advances in science and technology and the widespread adoption of computing have facilitated the continuous evolution of BPS technologies and mathematical algorithms, thereby rendering related research more achievable.
Most research has predominantly been conducted in Europe. Achieving the CER target by 2050 necessitates the extensive decarbonization retrofitting of large-scale existing buildings [168]. Moreover, with the fixed form and orientation of buildings slated for retrofitting, emphasis can be placed on optimizing the envelope and systems, thereby facilitating the application of optimization techniques [21].
Building CER projects typically involve multiple objectives and variables. Integrating BPS with data-driven methods such as optimization algorithms and ML can streamline the process, minimizing redundant tasks and facilitating a quicker and more accurate identification of optimal solutions.

5.2. Optimization and Surrogate Models

Using physical simulation methods to establish an optimization model can accurately reflect the relationship between building parameters and performance and evaluate the performance improvement potential, but the simulation time is relatively long. The use of surrogate models significantly enhances computational speed, and there exist variations in the selection and training of mathematical models across studies. For instance, the number of nodes in the hidden layer of an ANN varies depending on task objectives [169]. Likewise, the division ratio of the datasets into training and testing sets lacks uniformity, and the result reliability may hinge on the training and testing data used in surrogate model developments [135]. Therefore, the choice of model and computational approach should be determined by each task’s unique characteristics, such as the data volume and prediction goals [170].
Additionally, addressing issues such as model generalizations and transferability is needed to foster the widespread adoption of surrogate models. Transfer learning methods offer a viable approach to applying data acquired from previous tasks to similar or related tasks [171], which is particularly effective when the training dataset for the targeted task is not large [172]. The successful application of this approach depends on the similarity of data sources and the feasibility for predicting the building performance [173]. Similarly, in energy prediction research for large-scale buildings, transfer learning can solve problems like data scarcity and time-consuming data collection, enhancing the accuracy of data-driven models [174]. A study used transfer learning methods to apply evaluation models pre-trained using Swedish building datasets to predict the performance of buildings in China, which can effectively support building retrofit decisions in data-scarce regions [175].
The clarification and classification of optimization objectives and parameters can provide researchers with experience and paradigms for retrofits and serve as a reference for sensitivity analyses and variable screening outcomes. While focusing on CERs, the relevant studies also take into account optimization goals in terms of building ECs, thermal comfort, and costs [25]. A study also found that cost parameters are particularly common in MOO problems of private housing retrofits, and householders are more willing to carry out CER retrofits on the premise of economic benefits. Therefore, cost savings are as important as CERs, or even more critical [176]. Thus, achieving a balance among multiple performance indicators is essential in retrofitting efforts.
Retrofit strategies typically consider enhancing the performance of building envelopes and adjusting parameters of energy systems. Currently, there is a heightened focus on utilizing renewable energy sources, particularly PV and solar thermal systems. Additionally, emerging technologies and materials, such as switchable blinds [177], PV walls [178], electrochromic windows [179], phase change materials (PCMs), and cool paints [180], among others, are increasingly being employed to enhance the building performance. The vertical greening system (VGS) on roofs and walls are also noteworthy for their potential to mitigate GHGEs, urban heat islands, and noise pollution [181].

5.3. Optimization Methods and Tools

In building CER problems, the choice of optimization methods varies depending on specific objectives and parameters, necessitating tailored selections based on the actual problem [182]. Furthermore, it is crucial to assess the computational efficiency and stability of these methods. While prioritizing computational speed, it is essential to strike a balance between model simplicity and accuracy [97].
The utilization of general optimization platforms like MATLAB necessitates proficiency in a complex programming language, and the constant switching between modeling and optimization environments also poses an inconvenience. Hence, there is a need to integrate programming languages into toolkits that allow for direct applications. The Grasshopper-based Octopus tool [183] integrates optimization algorithms with building parametric modeling and is widely adopted for current BPS and optimization [184,185,186]. Nevertheless, further exploration is required, particularly in its application to building retrofit processes aimed at achieving CER goals. In addition, regarding the assessment of the solution set quality, a study compared the convergence and diversity using metrics such as IGD and HI [187].
In recent years, scholars have started to focus on integrating optimization algorithms with Building Information Modeling (BIM) to enhance decision-making abilities in building retrofitting [188]. As data-driven methods advance, there is a corresponding need to further develop tools that integrate these methods with simulation software, enabling timely and efficient applications in BPO. Additionally, establishing a database using existing building performance datasets [189] and implementing a decision support system based on data-driven methods can assist decision-makers in selecting optimal retrofit strategies based on real-world conditions [26].

5.4. Future Work

Uncertainties stemming from user behaviors and climate change present formidable challenges in building retrofitting [8]. Factors such as user preferences [190], comfort requirements [191], and energy-saving awareness exert a substantial influence on the building performance, making user behaviors one of the primary contributors to the disparity between BPS and actual operation data [192] and a pivotal factor shaping retrofit strategies [193]. For instance, Li et al. [194] integrated three user-behavior patterns—adjusting sunshade components, managing lighting, and ventilating by opening windows—into the optimization of sunshade devices. Their findings revealed how adaptive user behaviors significantly impact room illumination and the simulation of energy loads, thereby influencing the layout and sizing of sunshade panels. Only one of the reviewed studies discussed the impacts of user behaviors on building CEs [70]. This study focused on public buildings in the hot summer and cold winter region of China and implemented a time-of-use pricing strategy to regulate occupants’ energy-use behaviors, including the temperature set point of air conditioners and the usage time of electrical appliances. This strategy can reduce users’ ECs and CEs by 26.4% and energy expenditure by 26.1%. Current simulation studies often rely on fixed schedules, overlooking active user adjustments and the variability and spontaneity of behaviors. Future studies should incorporate more precise and comprehensive behavioral models into BPS frameworks [142].
A study developed a climate model based on atmospheric CO2 concentration changes, projecting that the Earth’s surface annual average temperature will rise by 1.5 °C by 2050 and by 2–4 °C by 2080 [195]. Given the long lifespan of buildings, climate change should be factored into energy-saving analyses [196] to ensure the applicability and efficacy of retrofit strategies [197]. Data-driven approaches can be used to generate future weather data and establish dynamic climate models.
The building performance is closely correlated with the outdoor environment [198]. Factors such as community greening, urban heat islands, and global warming alter microclimates, influencing the regional building performance [199]. Therefore, using uniform meteorological parameters for predicting building ECs can lead to inaccuracies [200], highlighting the need to integrate microclimate models into energy-efficiency predictions. Currently, optimization research for building CER retrofits has expanded beyond individual buildings to include local outdoor environments, exploring the effects of tree shading and the PV coverage of building façades on the life-cycle GWP of buildings and outdoor thermal comfort [201]. As such, there is still great potential for research on decarbonization retrofits at the scale of building groups or blocks.

6. Conclusions

This paper provides a review of the application of data-driven methods in building retrofits and performance optimization from the perspective of CERs and also discusses recent trends and advancements in various studies. A comparison of modeling methods, optimization variables, objectives, and decision-making methods across 45 relevant papers revealed the following findings:
  • There are usually two workflows to optimize the building performance. One is the workflow of the optimization of the physical simulation (model surrogate) performance: Using the combined input of a building site and energy-carbon-related retrofit variables, a BPO process based on a physical simulation is established. The generated datasets can be either iteratively processed with optimization algorithms directly or trained as a surrogate model, validated, and then processed using the MOO method. The other is the workflow of mathematical modeling-optimization analyses: with sufficient actual field-measured empirical data available, data-driven methods, such as regression or machine learning, are used to develop mathematical models, and multiple objectives are comprehensively optimized from the perspective of building CERs.
  • A building retrofit aims to maximize its benefits by integrating environmental, economic, and social considerations. Therefore, alongside CE objectives, factors like costs and thermal comfort should also be taken into account. There are 27 relevant studies in Table 2 related to the comprehensive optimization of three or more objectives, accounting for 60% of the total. Discussions on retrofit parameters should extensively cover aspects such as the thermal performance of the building envelope, building equipment and energy systems, and the utilization of renewable energy sources.
  • Data-driven methods applied in optimization enable the sampling, screening, and iterative refinement of retrofit plans using computational tools, facilitating the determination of optimal solutions. The advancement and deployment of surrogate models make simplified mathematical calculations replace complex physical simulations, which further enhance optimization efficiency while ensuring accuracy.
  • In the reviewed studies, only 2.2% (1 article) and 6.7% (3 articles) of the total focus on the impacts of human behaviors and climate change on building retrofits, respectively. Future research should delve deeper into the application of data-driven methods in building CER retrofits and BPO, considering user behaviors and variations in retrofit conditions amid long-term climate change scenarios. In addition, more work is needed to improve the accuracy of surrogate models and enhance generalizations and transfer capabilities.
Within the broader context of global warming and building stock renewals, building decarbonizing retrofitting and performance enhancements have emerged as focal points of research. This review highlights the capability of integrating data-driven methods with BPS to screen and determine retrofit plans, showing its significant applicational value in building CERs. It will also make a greater contribution to the CER targets of the building construction industry in the future.

Author Contributions

Conceptualization, S.-L.L. and F.Y.; methodology, S.-L.L., X.S. and F.Y.; software, S.-L.L.; formal analysis, S.-L.L., X.S. and F.Y.; investigation, S.-L.L.; resources, F.Y. and X.S.; data curation, S.-L.L.; writing—original draft preparation, S.-L.L.; writing—review and editing, F.Y. and X.S.; visualization, S.-L.L.; supervision, F.Y.; project administration, F.Y.; funding acquisition, F.Y. All authors have read and agreed to the published version of the manuscript.

Funding

This research was funded by the National Natural Science Foundation of China, grant number 52338004.

Data Availability Statement

No new data were created in this study.

Conflicts of Interest

The authors declare no conflicts of interest.

Abbreviations

ACAnnual cost
ACOAnt colony optimization
AHPAnalytic hierarchy process
ANNArtificial neural network
ASHRAEAmerican Society of Heating, Refrigerating, and Air-Conditioning Engineers
BPOBuilding performance optimization
BPSBuilding performance simulation
CEsCarbon emissions
CERCarbon emission reduction
CV(RMSE)Coefficient of variation of root mean square error
DNNDeep neural network
DOEsDesign of experiments
EAEvolutionary algorithm
ECEnergy consumption
ECEsEmbodied carbon emissions
EWSOAEnhanced water strider optimization algorithm
FEMPFederal Energy Management Program
GAGenetic algorithm
GBRTGradient boosting regression tree
GCGlobal cost
GHGEsGreenhouse gas emissions
GSAGlobal sensitivity analysis
GWPGlobal warming potential
IPMVPInternational Performance Measurement and Verification Protocol
LCALife-cycle assessment
LCCLife-cycle cost
LCCELife-cycle carbon emission
LCELife-cycle energy consumption
LSALocal sensitivity analysis
MAEMean absolute error
MAPEMean absolute percent error
MARSsMultiple adaptive regression splines
MBEMean bias error
MILPMixed-integer linear programming
MOEA/DMulti-objective evolutionary algorithm based on decomposition
MOGAMulti-objective genetic algorithm
MOOMulti-objective optimization
MVLRMulti-variate linear regression
NOPNonlinear optimization programming
NSGA-II/IIINon-dominated sorting genetic algorithm II/III
a/pNSGA-IIActive/passive archive NSGA-II
prNSGA-IIINSGA-III algorithm augmented by parallel computing structure and result-saving archive
OCOperational cost
OCESOperational carbon emissions
PRISMAPreferred reporting items for systemic reviews and meta-analyses
PSOParticle swarm optimization
RCRetrofit cost
SEGAStrengthen elitist genetic algorithm
SPEA2Strength Pareto evolutionary algorithm2
SQOLSocial quality of life
TDHSThermal discomfort hours
WCWater consumption

References

  1. Gan, V.J.L.; Deng, M.; Tse, K.T.; Chan, C.M.; Lo, I.M.C.; Cheng, J.C.P. Holistic BIM Framework for Sustainable Low Carbon Design of High-Rise Buildings. J. Clean. Prod. 2018, 195, 1091–1104. [Google Scholar] [CrossRef]
  2. Qian, D.; Li, Y.; Niu, F.; O’Neill, Z. Nationwide Savings Analysis of Energy Conservation Measures in Buildings. Energy Convers. Manag. 2019, 188, 1–18. [Google Scholar] [CrossRef]
  3. CO2 and Greenhouse Gas Emissions. Available online: https://ourworldindata.org/co2-and-greenhouse-gas-emissions (accessed on 5 March 2024).
  4. Mata, É.; Korpal, A.K.; Cheng, S.H.; Jiménez Navarro, J.P.; Filippidou, F.; Reyna, J.; Wang, R. A Map of Roadmaps for Zero and Low Energy and Carbon Buildings Worldwide. Environ. Res. Lett. 2020, 15, 113003. [Google Scholar] [CrossRef]
  5. Wei, Y.; Zhang, X.; Shi, Y.; Xia, L.; Pan, S.; Wu, J.; Han, M.; Zhao, X. A Review of Data-Driven Approaches for Prediction and Classification of Building Energy Consumption. Renew. Sustain. Energy Rev. 2018, 82, 1027–1047. [Google Scholar] [CrossRef]
  6. Love, P.; Arthur Bullen, P. Toward the Sustainable Adaptation of Existing Facilities. Facilities 2009, 27, 357–367. [Google Scholar] [CrossRef]
  7. Deb, C.; Schlueter, A. Review of Data-Driven Energy Modelling Techniques for Building Retrofit. Renew. Sustain. Energy Rev. 2021, 144, 110990. [Google Scholar] [CrossRef]
  8. Ma, Z.; Cooper, P.; Daly, D.; Ledo, L. Existing Building Retrofits: Methodology and State-of-the-Art. Energy Build. 2012, 55, 889–902. [Google Scholar] [CrossRef]
  9. Jensen, P.A.; Maslesa, E.; Berg, J.B.; Thuesen, C. 10 Questions Concerning Sustainable Building Renovation. Build. Environ. 2018, 143, 130–137. [Google Scholar] [CrossRef]
  10. Huang, J.; Wang, S.; Teng, F.; Feng, W. Thermal Performance Optimization of Envelope in the Energy-Saving Renovation of Existing Residential Buildings. Energy Build. 2021, 247, 111103. [Google Scholar] [CrossRef]
  11. Vilches, A.; Garcia-Martinez, A.; Sanchez-Montañes, B. Life Cycle Assessment (LCA) of Building Refurbishment: A Literature Review. Energy Build. 2017, 135, 286–301. [Google Scholar] [CrossRef]
  12. Ji, Y.; Xu, P.; Ye, Y. HVAC Terminal Hourly End-Use Disaggregation in Commercial Buildings with Fourier Series Model. Energy Build. 2015, 97, 33–46. [Google Scholar] [CrossRef]
  13. Jagarajan, R.; Abdullah Mohd Asmoni, M.N.; Mohammed, A.H.; Jaafar, M.N.; Lee Yim Mei, J.; Baba, M. Green Retrofitting—A Review of Current Status, Implementations and Challenges. Renew. Sustain. Energy Rev. 2017, 67, 1360–1368. [Google Scholar] [CrossRef]
  14. Nielsen, A.N.; Jensen, R.L.; Larsen, T.S.; Nissen, S.B. Early Stage Decision Support for Sustainable Building Renovation—A Review. Build. Environ. 2016, 103, 165–181. [Google Scholar] [CrossRef]
  15. De Wilde, P. Ten Questions Concerning Building Performance Analysis. Build. Environ. 2019, 153, 110–117. [Google Scholar] [CrossRef]
  16. Roman, N.D.; Bre, F.; Fachinotti, V.D.; Lamberts, R. Application and Characterization of Metamodels Based on Artificial Neural Networks for Building Performance Simulation: A Systematic Review. Energy Build. 2020, 217, 109972. [Google Scholar] [CrossRef]
  17. Wu, Z.; Xia, X.; Wang, B. Improving Building Energy Efficiency by Multiobjective Neighborhood Field Optimization. Energy Build. 2015, 87, 45–56. [Google Scholar] [CrossRef]
  18. Hong, T.; Langevin, J.; Sun, K. Building Simulation: Ten Challenges. Build. Simul. 2018, 11, 871–898. [Google Scholar] [CrossRef]
  19. Attia, S.; Hamdy, M.; O’Brien, W.; Carlucci, S. Assessing Gaps and Needs for Integrating Building Performance Optimization Tools in Net Zero Energy Buildings Design. Energy Build. 2013, 60, 110–124. [Google Scholar] [CrossRef]
  20. Di Biccari, C.; Calcerano, F.; D’Uffizi, F.; Esposito, A.; Campari, M.; Gigliarelli, E. Building Information Modeling and Building Performance Simulation Interoperability: State-of-the-Art and Trends in Current Literature. Adv. Eng. Inform. 2022, 54, 101753. [Google Scholar] [CrossRef]
  21. Shi, X.; Tian, Z.; Chen, W.; Si, B.; Jin, X. A Review on Building Energy Efficient Design Optimization Rom the Perspective of Architects. Renew. Sustain. Energy Rev. 2016, 65, 872–884. [Google Scholar] [CrossRef]
  22. Westermann, P.; Evins, R. Surrogate Modelling for Sustainable Building Design—A Review. Energy Build. 2019, 198, 170–186. [Google Scholar] [CrossRef]
  23. Luo, X.J.; Oyedele, L.O. A Data-Driven Life-Cycle Optimisation Approach for Building Retrofitting: A Comprehensive Assessment on Economy, Energy and Environment. J. Build. Eng. 2021, 43, 102934. [Google Scholar] [CrossRef]
  24. Thrampoulidis, E.; Mavromatidis, G.; Lucchi, A.; Orehounig, K. A Machine Learning-Based Surrogate Model to Approximate Optimal Building Retrofit Solutions. Appl. Energy 2021, 281, 116024. [Google Scholar] [CrossRef]
  25. Ascione, F.; Bianco, N.; Mauro, G.; Napolitano, D.; Vanoli, G. A Multi-Criteria Approach to Achieve Constrained Cost-Optimal Energy Retrofits of Buildings by Mitigating Climate Change and Urban Overheating. Climate 2018, 6, 37. [Google Scholar] [CrossRef]
  26. Liu, Z.; Yu, C.; Qian, Q.K.; Huang, R.; You, K.; Visscher, H.; Zhang, G. Incentive Initiatives on Energy-Efficient Renovation of Existing Buildings towards Carbon–Neutral Blueprints in China: Advancements, Challenges and Prospects. Energy Build. 2023, 296, 113343. [Google Scholar] [CrossRef]
  27. Li, Y.; Du, H.; Kumaraswamy, S.B. Case-Based Reasoning Approach for Decision-Making in Building Retrofit: A Review. Build. Environ. 2024, 248, 111030. [Google Scholar] [CrossRef]
  28. Mostafavi, F.; Tahsildoost, M.; Zomorodian, Z. Energy Efficiency and Carbon Emission in High-Rise Buildings: A Review (2005–2020). Build. Environ. 2021, 206, 108329. [Google Scholar] [CrossRef]
  29. Pombo, O.; Rivela, B.; Neila, J. The Challenge of Sustainable Building Renovation: Assessment of Current Criteria and Future Outlook. J. Clean. Prod. 2016, 123, 88–100. [Google Scholar] [CrossRef]
  30. Li, J.; Ng, S.T.; Skitmore, M. Review of Low-Carbon Refurbishment Solutions for Residential Buildings with Particular Reference to Multi-Story Buildings in Hong Kong. Renew. Sustain. Energy Rev. 2017, 73, 393–407. [Google Scholar] [CrossRef]
  31. Aghamolaei, R.; Fallahpour, M. Strategies towards Reducing Carbon Emission in University Campuses: A Comprehensive Review of Both Global and Local Scales. J. Build. Eng. 2023, 76, 107183. [Google Scholar] [CrossRef]
  32. Sun, Y.; Haghighat, F.; Fung, B.C.M. A Review of The-State-of-the-Art in Data-Driven Approaches for Building Energy Prediction. Energy Build. 2020, 221, 110022. [Google Scholar] [CrossRef]
  33. Page, M.J.; McKenzie, J.E.; Bossuyt, P.M.; Boutron, I.; Hoffmann, T.C.; Mulrow, C.D.; Shamseer, L.; Tetzlaff, J.M.; Akl, E.A.; Brennan, S.E.; et al. The PRISMA 2020 Statement: An Updated Guideline for Reporting Systematic Reviews. Syst. Rev. 2021, 10, 89. [Google Scholar] [CrossRef] [PubMed]
  34. Son, T.H.; Weedon, Z.; Yigitcanlar, T.; Sanchez, T.; Corchado, J.M.; Mehmood, R. Algorithmic Urban Planning for Smart and Sustainable Development: Systematic Review of the Literature. Sustain. Cities Soc. 2023, 94, 104562. [Google Scholar] [CrossRef]
  35. Mongeon, P.; Paul-Hus, A. The Journal Coverage of Web of Science and Scopus: A Comparative Analysis. Scientometrics 2016, 106, 213–228. [Google Scholar] [CrossRef]
  36. Olawumi, T.O.; Chan, D.W.M. A Scientometric Review of Global Research on Sustainability and Sustainable Development. J. Clean. Prod. 2018, 183, 231–250. [Google Scholar] [CrossRef]
  37. Santos, R.; Costa, A.A.; Grilo, A. Bibliometric Analysis and Review of Building Information Modelling Literature Published between 2005 and 2015. Autom. Constr. 2017, 80, 118–136. [Google Scholar] [CrossRef]
  38. Luo, X.J. An Integrated Passive and Active Retrofitting Approach toward Minimum Whole-Life Carbon Footprint. Energy Build. 2023, 295, 113337. [Google Scholar] [CrossRef]
  39. Luo, X.J.; Oyedele, L.O. Assessment and Optimisation of Life Cycle Environment, Economy and Energy for Building Retrofitting. Energy Sustain. Dev. 2021, 65, 77–100. [Google Scholar] [CrossRef]
  40. Liang, Y.; Pan, Y.; Yuan, X.; Yang, Y.; Fu, L.; Li, J.; Sun, T.; Huang, Z.; Kosonen, R. Assessment of Operational Carbon Emission Reduction of Energy Conservation Measures for Commercial Buildings: Model Development. Energy Build. 2022, 268, 112189. [Google Scholar] [CrossRef]
  41. Aghamolaei, R.; Ghaani, M.R. Balancing the Impacts of Energy Efficiency Strategies on Comfort Quality of Interior Places: Application of Optimization Algorithms in Domestic Housing. J. Build. Eng. 2020, 29, 101174. [Google Scholar] [CrossRef]
  42. Charles, A.; Maref, W.; Ouellet-Plamondon, C.M. Case Study of the Upgrade of an Existing Office Building for Low Energy Consumption and Low Carbon Emissions. Energy Build. 2019, 183, 151–160. [Google Scholar] [CrossRef]
  43. Niemelä, T.; Levy, K.; Kosonen, R.; Jokisalo, J. Cost-Optimal Renovation Solutions to Maximize Environmental Performance, Indoor Thermal Conditions and Productivity of Office Buildings in Cold Climate. Sustain. Cities Soc. 2017, 32, 417–434. [Google Scholar] [CrossRef]
  44. Amani, N.; Kiaee, E. Developing a Two-Criteria Framework to Rank Thermal Insulation Materials in Nearly Zero Energy Buildings Using Multi-Objective Optimization Approach. J. Clean. Prod. 2020, 276, 122592. [Google Scholar] [CrossRef]
  45. Sharif, S.A.; Hammad, A. Developing Surrogate ANN for Selecting Near-Optimal Building Energy Renovation Methods Considering Energy Consumption, LCC and LCA. J. Build. Eng. 2019, 25, 100790. [Google Scholar] [CrossRef]
  46. Jeong, K.; Hong, T.; Kim, J.; Cho, K. Development of a Multi-Objective Optimization Model for Determining the Optimal CO2 Emissions Reduction Strategies for a Multi-Family Housing Complex. Renew. Sustain. Energy Rev. 2019, 110, 118–131. [Google Scholar] [CrossRef]
  47. Chantrelle, F.P.; Lahmidi, H.; Keilholz, W.; Mankibi, M.E.; Michel, P. Development of a Multicriteria Tool for Optimizing the Renovation of Buildings. Appl. Energy 2011, 88, 1386–1394. [Google Scholar] [CrossRef]
  48. Salata, F.; Ciancio, V.; Dell’Olmo, J.; Golasi, I.; Palusci, O.; Coppi, M. Effects of Local Conditions on the Multi-Variable and Multi-Objective Energy Optimization of Residential Buildings Using Genetic Algorithms. Appl. Energy 2020, 260, 114289. [Google Scholar] [CrossRef]
  49. Niemelä, T.; Kosonen, R.; Jokisalo, J. Energy Performance and Environmental Impact Analysis of Cost-Optimal Renovation Solutions of Large Panel Apartment Buildings in Finland. Sustain. Cities Soc. 2017, 32, 9–30. [Google Scholar] [CrossRef]
  50. Mostafazadeh, F.; Eirdmousa, S.J.; Tavakolan, M. Energy, Economic and Comfort Optimization of Building Retrofits Considering Climate Change: A Simulation-Based NSGA-III Approach. Energy Build. 2023, 280, 112721. [Google Scholar] [CrossRef]
  51. Gao, Y.; Luo, S.; Jiang, J.; Rong, Y. Environmental-Thermal-Economic Performance Trade-off for Rural Residence Retrofitting in the Beijing–Tianjin–Hebei Region, Northern China: A Multi-Objective Optimisation Framework under Different Scenarios. Energy Build. 2023, 286, 112910. [Google Scholar] [CrossRef]
  52. Li, Q.; Hu, H.; Ma, L.; Wang, Z.; Arıcı, M.; Li, D.; Luo, D.; Jia, J.; Jiang, W.; Qi, H. Evaluation of Energy-Saving Retrofits for Sunspace of Rural Residential Buildings Based on Orthogonal Experiment and Entropy Weight Method. Energy Sustain. Dev. 2022, 70, 569–580. [Google Scholar] [CrossRef]
  53. Kertsmik, K.-A.; Kuusk, K.; Lylykangas, K.; Kalamees, T. Evaluation of Renovation Strategies: Cost-Optimal, CO2e Optimal, or Total Energy Optimal? Energy Build. 2023, 287, 112995. [Google Scholar] [CrossRef]
  54. Son, H.; Kim, C. Evolutionary Many-Objective Optimization for Retrofit Planning in Public Buildings: A Comparative Study. J. Clean. Prod. 2018, 190, 403–410. [Google Scholar] [CrossRef]
  55. Song, J.; Wang, W.; Ni, P.; Zheng, H.; Zhang, Z.; Zhou, Y. Framework on Low-Carbon Retrofit of Rural Residential Buildings in Arid Areas of Northwest China: A Case Study of Turpan Residential Buildings. Build. Simul. 2023, 16, 279–297. [Google Scholar] [CrossRef]
  56. Liu, S.; Wang, Y.; Liu, X.; Yang, L.; Zhang, Y.; He, J. How Does Future Climatic Uncertainty Affect Multi-Objective Building Energy Retrofit Decisions? Evidence from Residential Buildings in Subtropical Hong Kong. Sustain. Cities Soc. 2023, 92, 104482. [Google Scholar] [CrossRef]
  57. Schwartz, Y.; Raslan, R.; Mumovic, D. Implementing Multi Objective Genetic Algorithm for Life Cycle Carbon Footprint and Life Cycle Cost Minimisation: A Building Refurbishment Case Study. Energy 2016, 97, 58–68. [Google Scholar] [CrossRef]
  58. Shadram, F.; Mukkavaara, J. Improving Life Cycle Sustainability and Profitability of Buildings through Optimization: A Case Study. Buildings 2022, 12, 497. [Google Scholar] [CrossRef]
  59. Miglani, S.; Orehounig, K.; Carmeliet, J. Integrating a Thermal Model of Ground Source Heat Pumps and Solar Regeneration within Building Energy System Optimization. Appl. Energy 2018, 218, 78–94. [Google Scholar] [CrossRef]
  60. Zhang, H.; Hewage, K.; Prabatha, T.; Sadiq, R. Life Cycle Thinking-Based Energy Retrofits Evaluation Framework for Canadian Residences: A Pareto Optimization Approach. Build. Environ. 2021, 204, 108115. [Google Scholar] [CrossRef]
  61. Seyedzadeh, S.; Pour Rahimian, F.; Oliver, S.; Rodriguez, S.; Glesk, I. Machine Learning Modelling for Predicting Non-Domestic Buildings Energy Performance: A Model to Support Deep Energy Retrofit Decision-Making. Appl. Energy 2020, 279, 115908. [Google Scholar] [CrossRef]
  62. Ciardiello, A.; Rosso, F.; Dell’Olmo, J.; Ciancio, V.; Ferrero, M.; Salata, F. Multi-Objective Approach to the Optimization of Shape and Envelope in Building Energy Design. Appl. Energy 2020, 280, 115984. [Google Scholar] [CrossRef]
  63. Liu, B.; Pouramini, S. Multi-Objective Optimization for Thermal Comfort Enhancement and Greenhouse Gas Emission Reduction in Residential Buildings Applying Retrofitting Measures by an Enhanced Water Strider Optimization Algorithm: A Case Study. Energy Rep. 2021, 7, 1915–1929. [Google Scholar] [CrossRef]
  64. Montana, F.; Kanafani, K.; Wittchen, K.; Birgisdottir, H.; Longo, S.; Cellura, M.; Riva Sanseverino, E. Multi-Objective Optimization of Building Life Cycle Performance. A Housing Renovation Case Study in Northern Europe. Sustainability 2020, 12, 7807. [Google Scholar] [CrossRef]
  65. Rosso, F.; Ciancio, V.; Dell’Olmo, J.; Salata, F. Multi-Objective Optimization of Building Retrofit in the Mediterranean Climate by Means of Genetic Algorithm Application. Energy Build. 2020, 216, 109945. [Google Scholar] [CrossRef]
  66. Kadrić, D.; Aganović, A.; Kadrić, E. Multi-Objective Optimization of Energy-Efficient Retrofitting Strategies for Single-Family Residential Homes: Minimizing Energy Consumption, CO2 Emissions and Retrofit Costs. Energy Rep. 2023, 10, 1968–1981. [Google Scholar] [CrossRef]
  67. Gao, B.; Zhu, X.; Ren, J.; Ran, J.; Kim, M.K.; Liu, J. Multi-Objective Optimization of Energy-Saving Measures and Operation Parameters for a Newly Retrofitted Building in Future Climate Conditions: A Case Study of an Office Building in Chengdu. Energy Rep. 2023, 9, 2269–2285. [Google Scholar] [CrossRef]
  68. Wu, R.; Mavromatidis, G.; Orehounig, K.; Carmeliet, J. Multiobjective Optimisation of Energy Systems and Building Envelope Retrofit in a Residential Community. Appl. Energy 2017, 190, 634–649. [Google Scholar] [CrossRef]
  69. Schütz, T.; Schiffer, L.; Harb, H.; Fuchs, M.; Müller, D. Optimal Design of Energy Conversion Units and Envelopes for Residential Building Retrofits Using a Comprehensive MILP Model. Appl. Energy 2017, 185, 1–15. [Google Scholar] [CrossRef]
  70. Huang, H.; Wang, H.; Hu, Y.-J.; Li, C.; Wang, X. Optimal Plan for Energy Conservation and CO2 Emissions Reduction of Public Buildings Considering Users’ Behavior: Case of China. Energy 2022, 261, 125037. [Google Scholar] [CrossRef]
  71. Panagiotidou, M.; Aye, L.; Rismanchi, B. Optimisation of Multi-Residential Building Retrofit, Cost-Optimal and Net-Zero Emission Targets. Energy Build. 2021, 252, 111385. [Google Scholar] [CrossRef]
  72. Dou, Z.; Jin, L.; Chen, Y.; Huang, Z. Optimization of Cost–Carbon Reduction–Technology Solution for Existing Office Parks Based on Genetic Algorithm. Processes 2023, 11, 2452. [Google Scholar] [CrossRef]
  73. Xu, Y.; Yan, C.; Wang, G.; Shi, J.; Sheng, K.; Li, J.; Jiang, Y. Optimization Research on Energy-Saving and Life-Cycle Decarbonization Retrofitting of Existing School Buildings: A Case Study of a School in Nanjing. Sol. Energy 2023, 254, 54–66. [Google Scholar] [CrossRef]
  74. Karatas, A.; El-Rayes, K. Optimizing Tradeoffs among Housing Sustainability Objectives. Autom. Constr. 2015, 53, 83–94. [Google Scholar] [CrossRef]
  75. Schwartz, Y.; Raslan, R.; Mumovic, D. Refurbish or Replace? The Life Cycle Carbon Footprint and Life Cycle Cost of Refurbished and New Residential Archetype Buildings in London. Energy 2022, 248, 123585. [Google Scholar] [CrossRef]
  76. Sharif, S.A.; Hammad, A. Simulation-Based Multi-Objective Optimization of Institutional Building Renovation Considering Energy Consumption, Life-Cycle Cost and Life-Cycle Assessment. J. Build. Eng. 2019, 21, 429–445. [Google Scholar] [CrossRef]
  77. Saad, M.M.; Menon, R.P.; Eicker, U. Supporting Decision Making for Building Decarbonization: Developing Surrogate Models for Multi-Criteria Building Retrofitting Analysis. Energies 2023, 16, 6030. [Google Scholar] [CrossRef]
  78. Wang, P.; Gong, G.; Wang, Y.; Li, L. Thermodynamic Investigation of Building Integrated Energy Efficiency for Building Retrofit. Energy Build. 2014, 77, 139–148. [Google Scholar] [CrossRef]
  79. Galimshina, A.; Moustapha, M.; Hollberg, A.; Padey, P.; Lasvaux, S.; Sudret, B.; Habert, G. What Is the Optimal Robust Environmental and Cost-Effective Solution for Building Renovation? Not the Usual One. Energy Build. 2021, 251, 111329. [Google Scholar] [CrossRef]
  80. Mei, J.; Xu, G.; Ahmad, W.; Khan, K.; Amin, M.N.; Aslam, F.; Alaskar, A. Promoting Sustainable Materials Using Recycled Rubber in Concrete: A Review. J. Clean. Prod. 2022, 373, 133927. [Google Scholar] [CrossRef]
  81. De Boeck, L.; Verbeke, S.; Audenaert, A.; De Mesmaeker, L. Improving the Energy Performance of Residential Buildings: A Literature Review. Renew. Sustain. Energy Rev. 2015, 52, 960–975. [Google Scholar] [CrossRef]
  82. Shaamala, A.; Yigitcanlar, T.; Nili, A.; Nyandega, D. Algorithmic Green Infrastructure Optimisation: Review of Artificial Intelligence Driven Approaches for Tackling Climate Change. Sustain. Cities Soc. 2024, 101, 105182. [Google Scholar] [CrossRef]
  83. Ellegaard, O.; Wallin, J.A. The Bibliometric Analysis of Scholarly Production: How Great Is the Impact? Scientometrics 2015, 105, 1809–1831. [Google Scholar] [CrossRef] [PubMed]
  84. Cobo, M.J.; López-Herrera, A.G.; Herrera-Viedma, E.; Herrera, F. Science Mapping Software Tools: Review, Analysis, and Cooperative Study among Tools. J. Am. Soc. Inf. Sci. Technol. 2011, 62, 1382–1402. [Google Scholar] [CrossRef]
  85. Chen, C. CiteSpace II: Detecting and Visualizing Emerging Trends and Transient Patterns in Scientific Literature. J. Am. Soc. Inf. Sci. Technol. 2006, 57, 359–377. [Google Scholar] [CrossRef]
  86. Tetteh, M.O.; Darko, A.; Chan, A.P.C.; Jafari, A.; Brilakis, I.; Chen, W.; Nani, G.; Kwame Yevu, S. Scientometric Mapping of Global Research on Green Retrofitting of Existing Buildings (GREB): Pathway towards a Holistic GREB Framework. Energy Build. 2022, 277, 112532. [Google Scholar] [CrossRef]
  87. Zhang, N.; Yang, C.; Wang, S. Research Progress and Prospect of Environmental, Social and Governance: A Systematic Literature Review and Bibliometric Analysis. J. Clean. Prod. 2024, 447, 141489. [Google Scholar] [CrossRef]
  88. Debrah, C.; Chan, A.P.C.; Darko, A. Artificial Intelligence in Green Building. Autom. Constr. 2022, 137, 104192. [Google Scholar] [CrossRef]
  89. Zhao, H.; Magoulès, F. A Review on the Prediction of Building Energy Consumption. Renew. Sustain. Energy Rev. 2012, 16, 3586–3592. [Google Scholar] [CrossRef]
  90. Shen, J.; Krietemeyer, B.; Bartosh, A.; Gao, Z.; Zhang, J. Green Design Studio: A Modular-Based Approach for High-Performance Building Design. Build. Simul. 2021, 14, 241–268. [Google Scholar] [CrossRef]
  91. DesignBuilder Software Ltd-Home. Available online: https://designbuilder.co.uk/ (accessed on 21 February 2024).
  92. OpenStudio. Available online: https://openstudio.net/ (accessed on 21 February 2024).
  93. 3D Design Software|3D Modeling on the Web. Available online: http://www.sketchup.com/page/homepage (accessed on 21 February 2024).
  94. Ladybug Tools|Honeybee. Available online: https://www.ladybug.tools/honeybee.html (accessed on 11 March 2024).
  95. Network, S.D. Created This N. Grasshopper. Available online: https://www.grasshopper3d.com/ (accessed on 21 February 2024).
  96. Li, C.Z.; Zhang, L.; Liang, X.; Xiao, B.; Tam, V.W.Y.; Lai, X.; Chen, Z. Advances in the Research of Building Energy Saving. Energy Build. 2022, 254, 111556. [Google Scholar] [CrossRef]
  97. Pan, Y.; Zhu, M.; Lv, Y.; Yang, Y.; Liang, Y.; Yin, R.; Yang, Y.; Jia, X.; Wang, X.; Zeng, F.; et al. Building Energy Simulation and Its Application for Building Performance Optimization: A Review of Methods, Tools, and Case Studies. Adv. Appl. Energy 2023, 10, 100135. [Google Scholar] [CrossRef]
  98. Guo, J.; Liu, R.; Xia, T.; Pouramini, S. Energy Model Calibration in an Office Building by an Optimization-Based Method. Energy Rep. 2021, 7, 4397–4411. [Google Scholar] [CrossRef]
  99. ASHRAE. Guideline 14, Measurement of Energy and Demand Savings; American Society of Heating, Refrigerating, and Air Conditioning Engineers: Atlanta, GA, USA, 2002. [Google Scholar]
  100. EVO. International Performance Measurement and Verification Protocol: Concepts and Options for Determining Energy and Water Savings; Efficiency Valuation Organization: Washington, DC, USA, 2002; Volume 1. [Google Scholar]
  101. US DOE FEMP. M&V Guidelines: Measurement and Verification for Performance-Based Contracts, Version 4.0; Energy Efficiency and Renewable Energy; US DOE FEMP: Washington, DC, USA, 2015.
  102. Hong, T.; Kim, J.; Koo, C. LCC and LCCO2 Analysis of Green Roofs in Elementary Schools with Energy Saving Measures. Energy Build. 2012, 45, 229–239. [Google Scholar] [CrossRef]
  103. Coakley, D.; Raftery, P.; Keane, M. A Review of Methods to Match Building Energy Simulation Models to Measured Data. Renew. Sustain. Energy Rev. 2014, 37, 123–141. [Google Scholar] [CrossRef]
  104. Nguyen, A.-T.; Reiter, S.; Rigo, P. A Review on Simulation-Based Optimization Methods Applied to Building Performance Analysis. Appl. Energy 2014, 113, 1043–1058. [Google Scholar] [CrossRef]
  105. Tian, W. A Review of Sensitivity Analysis Methods in Building Energy Analysis. Renew. Sustain. Energy Rev. 2013, 20, 411–419. [Google Scholar] [CrossRef]
  106. Hou, D.; Hassan, I.G.; Wang, L. Review on Building Energy Model Calibration by Bayesian Inference. Renew. Sustain. Energy Rev. 2021, 143, 110930. [Google Scholar] [CrossRef]
  107. Huang, H.; Wang, H.; Hu, Y.-J.; Li, C.; Wang, X. The Development Trends of Existing Building Energy Conservation and Emission Reduction—A Comprehensive Review. Energy Rep. 2022, 8, 13170–13188. [Google Scholar] [CrossRef]
  108. Torabi Moghadam, S.; Toniolo, J.; Mutani, G.; Lombardi, P. A GIS-Statistical Approach for Assessing Built Environment Energy Use at Urban Scale. Sustain. Cities Soc. 2018, 37, 70–84. [Google Scholar] [CrossRef]
  109. Yussuf, R.O.; Asfour, O.S. Applications of Artificial Intelligence for Energy Efficiency throughout the Building Lifecycle: An Overview. Energy Build. 2024, 305, 113903. [Google Scholar] [CrossRef]
  110. Seyedzadeh, S.; Rahimian, F.P.; Oliver, S.; Glesk, I.; Kumar, B. Data Driven Model Improved by Multi-Objective Optimisation for Prediction of Building Energy Loads. Autom. Constr. 2020, 116, 103188. [Google Scholar] [CrossRef]
  111. Grillone, B.; Danov, S.; Sumper, A.; Cipriano, J.; Mor, G. A Review of Deterministic and Data-Driven Methods to Quantify Energy Efficiency Savings and to Predict Retrofitting Scenarios in Buildings. Renew. Sustain. Energy Rev. 2020, 131, 110027. [Google Scholar] [CrossRef]
  112. Seyedzadeh, S.; Pour Rahimian, F.; Rastogi, P.; Glesk, I. Tuning Machine Learning Models for Prediction of Building Energy Loads. Sustain. Cities Soc. 2019, 47, 101484. [Google Scholar] [CrossRef]
  113. Magnier, L.; Haghighat, F. Multiobjective Optimization of Building Design Using TRNSYS Simulations, Genetic Algorithm, and Artificial Neural Network. Build. Environ. 2010, 45, 739–746. [Google Scholar] [CrossRef]
  114. Alwosheel, A.; Van Cranenburgh, S.; Chorus, C.G. Is Your Dataset Big Enough? Sample Size Requirements When Using Artificial Neural Networks for Discrete Choice Analysis. J. Choice Model. 2018, 28, 167–182. [Google Scholar] [CrossRef]
  115. Liu, C.; Sun, B.; Zhang, C.; Li, F. A Hybrid Prediction Model for Residential Electricity Consumption Using Holt-Winters and Extreme Learning Machine. Appl. Energy 2020, 275, 115383. [Google Scholar] [CrossRef]
  116. Van Gelder, L.; Das, P.; Janssen, H.; Roels, S. Comparative Study of Metamodelling Techniques in Building Energy Simulation: Guidelines for Practitioners. Simul. Model. Pract. Theory 2014, 49, 245–257. [Google Scholar] [CrossRef]
  117. Claesen, M.; De Moor, B. Hyperparameter Search in Machine Learning. arXiv 2015, arXiv:1502.02127. [Google Scholar]
  118. Rätz, M.; Javadi, A.P.; Baranski, M.; Finkbeiner, K.; Müller, D. Automated Data-Driven Modeling of Building Energy Systems via Machine Learning Algorithms. Energy Build. 2019, 202, 109384. [Google Scholar] [CrossRef]
  119. Chen, Z.; Chen, Y.; Xiao, T.; Wang, H.; Hou, P. A Novel Short-Term Load Forecasting Framework Based on Time-Series Clustering and Early Classification Algorithm. Energy Build. 2021, 251, 111375. [Google Scholar] [CrossRef]
  120. Amasyali, K.; El-Gohary, N.M. A Review of Data-Driven Building Energy Consumption Prediction Studies. Renew. Sustain. Energy Rev. 2018, 81, 1192–1205. [Google Scholar] [CrossRef]
  121. Al-Shargabi, A.A.; Almhafdy, A.; Ibrahim, D.M.; Alghieth, M.; Chiclana, F. Buildings’ Energy Consumption Prediction Models Based on Buildings’ Characteristics: Research Trends, Taxonomy, and Performance Measures. J. Build. Eng. 2022, 54, 104577. [Google Scholar] [CrossRef]
  122. Abbasabadi, N.; Ashayeri, M.; Azari, R.; Stephens, B.; Heidarinejad, M. An Integrated Data-Driven Framework for Urban Energy Use Modeling (UEUM). Appl. Energy 2019, 253, 113550. [Google Scholar] [CrossRef]
  123. Chen, Y.; Guo, M.; Chen, Z.; Chen, Z.; Ji, Y. Physical Energy and Data-Driven Models in Building Energy Prediction: A Review. Energy Rep. 2022, 8, 2656–2671. [Google Scholar] [CrossRef]
  124. Venkatraj, V.; Dixit, M.K. Challenges in Implementing Data-Driven Approaches for Building Life Cycle Energy Assessment: A Review. Renew. Sustain. Energy Rev. 2022, 160, 112327. [Google Scholar] [CrossRef]
  125. Fenner, A.E.; Kibert, C.J.; Li, J.; Razkenari, M.A.; Hakim, H.; Lu, X.; Kouhirostami, M.; Sam, M. Embodied, Operation, and Commuting Emissions: A Case Study Comparing the Carbon Hotspots of an Educational Building. J. Clean. Prod. 2020, 268, 122081. [Google Scholar] [CrossRef]
  126. Chen, T.Y.; Burnett, J.; Chau, C.K. Analysis of Embodied Energy Use in the Residential Building of Hong Kong. Energy 2001, 26, 323–340. [Google Scholar] [CrossRef]
  127. Elaouzy, Y.; El Fadar, A. Energy, Economic and Environmental Benefits of Integrating Passive Design Strategies into Buildings: A Review. Renew. Sustain. Energy Rev. 2022, 167, 112828. [Google Scholar] [CrossRef]
  128. Lin, M.; Ali, A.; Andargie, M.S.; Azar, E. Multidomain Drivers of Occupant Comfort, Productivity, and Well-Being in Buildings: Insights from an Exploratory and Explanatory Analysis. J. Manag. Eng. 2021, 37, 04021020. [Google Scholar] [CrossRef]
  129. He, C.; Hou, Y.; Ding, L.; Li, P. Visualized Literature Review on Sustainable Building Renovation. J. Build. Eng. 2021, 44, 102622. [Google Scholar] [CrossRef]
  130. Evins, R. A Review of Computational Optimisation Methods Applied to Sustainable Building Design. Renew. Sustain. Energy Rev. 2013, 22, 230–245. [Google Scholar] [CrossRef]
  131. Hashempour, N.; Taherkhani, R.; Mahdikhani, M. Energy Performance Optimization of Existing Buildings: A Literature Review. Sustain. Cities Soc. 2020, 54, 101967. [Google Scholar]
  132. Urbikain, M.K. Energy Efficient Solutions for Retrofitting a Residential Multi-Storey Building with Vacuum Insulation Panels and Low-E Windows in Two European Climates. J. Clean. Prod. 2020, 269, 121459. [Google Scholar] [CrossRef]
  133. Liu, C.; Sharples, S.; Mohammadpourkarbasi, H. A Review of Building Energy Retrofit Measures, Passive Design Strategies and Building Regulation for the Low Carbon Development of Existing Dwellings in the Hot Summer–Cold Winter Region of China. Energies 2023, 16, 4115. [Google Scholar] [CrossRef]
  134. Hamby, D.M. A Review of Techniques for Parameter Sensitivity Analysis of Environmental Models. Environ. Monit. Assess. 1994, 32, 135–154. [Google Scholar] [CrossRef]
  135. Kheiri, F. A Review on Optimization Methods Applied in Energy-Efficient Building Geometry and Envelope Design. Renew. Sustain. Energy Rev. 2018, 92, 897–920. [Google Scholar] [CrossRef]
  136. Heiselberg, P.; Brohus, H.; Hesselholt, A.; Rasmussen, H.; Seinre, E.; Thomas, S. Application of Sensitivity Analysis in Design of Sustainable Buildings. Renew. Energy 2009, 34, 2030–2036. [Google Scholar] [CrossRef]
  137. Tian, Z.; Zhang, X.; Jin, X.; Zhou, X.; Si, B.; Shi, X. Towards Adoption of Building Energy Simulation and Optimization for Passive Building Design: A Survey and a Review. Energy Build. 2018, 158, 1306–1316. [Google Scholar] [CrossRef]
  138. Asadi, E.; Silva, M.G.D.; Antunes, C.H.; Dias, L.; Glicksman, L. Multi-Objective Optimization for Building Retrofit: A Model Using Genetic Algorithm and Artificial Neural Network and an Application. Energy Build. 2014, 81, 444–456. [Google Scholar] [CrossRef]
  139. Tian, W.; Heo, Y.; De Wilde, P.; Li, Z.; Yan, D.; Park, C.S.; Feng, X.; Augenbroe, G. A Review of Uncertainty Analysis in Building Energy Assessment. Renew. Sustain. Energy Rev. 2018, 93, 285–301. [Google Scholar] [CrossRef]
  140. Kim, Y.-J.; Ahn, K.-U.; Park, C.-S. Decision Making of HVAC System Using Bayesian Markov Chain Monte Carlo Method. Energy Build. 2014, 72, 112–121. [Google Scholar] [CrossRef]
  141. Longo, S.; Montana, F.; Riva Sanseverino, E. A Review on Optimization and Cost-Optimal Methodologies in Low-Energy Buildings Design and Environmental Considerations. Sustain. Cities Soc. 2019, 45, 87–104. [Google Scholar] [CrossRef]
  142. Huang, Y.; Niu, J. Optimal Building Envelope Design Based on Simulated Performance: History, Current Status and New Potentials. Energy Build. 2016, 117, 387–398. [Google Scholar] [CrossRef]
  143. Deb, K.; Pratap, A.; Agarwal, S.; Meyarivan, T. A Fast and Elitist Multiobjective Genetic Algorithm: NSGA-II. IEEE Trans. Evol. Comput. 2002, 6, 182–197. [Google Scholar] [CrossRef]
  144. Liu, Y.; Heuvelink, G.B.M.; Bai, Z.; He, P.; Xu, X.; Ding, W.; Huang, S. Analysis of Spatio-Temporal Variation of Crop Yield in China Using Stepwise Multiple Linear Regression. Field Crops Res. 2021, 264, 108098. [Google Scholar] [CrossRef]
  145. Ali, A.; Jayaraman, R.; Azar, E.; Maalouf, M. A Comparative Analysis of Machine Learning and Statistical Methods for Evaluating Building Performance: A Systematic Review and Future Benchmarking Framework. Build. Environ. 2024, 252, 111268. [Google Scholar] [CrossRef]
  146. Amini Toosi, H.; Lavagna, M.; Leonforte, F.; Del Pero, C.; Aste, N. Life Cycle Sustainability Assessment in Building Energy Retrofitting; A Review. Sustain. Cities Soc. 2020, 60, 102248. [Google Scholar] [CrossRef]
  147. Mousavi, S.; Villarreal-Marroquín, M.G.; Hajiaghaei-Keshteli, M.; Smith, N.R. Data-Driven Prediction and Optimization toward Net-Zero and Positive-Energy Buildings: A Systematic Review. Build. Environ. 2023, 242, 110578. [Google Scholar] [CrossRef]
  148. Machairas, V.; Tsangrassoulis, A.; Axarli, K. Algorithms for Optimization of Building Design: A Review. Renew. Sustain. Energy Rev. 2014, 31, 101–112. [Google Scholar] [CrossRef]
  149. Congradac, V.; Kulic, F. HVAC System Optimization with CO2 Concentration Control Using Genetic Algorithms. Energy Build. 2009, 41, 571–577. [Google Scholar] [CrossRef]
  150. Delgarm, N.; Sajadi, B.; Delgarm, S.; Kowsary, F. A Novel Approach for the Simulation-Based Optimization of the Buildings Energy Consumption Using NSGA-II: Case Study in Iran. Energy Build. 2016, 127, 552–560. [Google Scholar] [CrossRef]
  151. Deb, K.; Thiele, L.; Laumanns, M.; Zitzler, E. Scalable Test Problems for Evolutionary Multiobjective Optimization. In Evolutionary Multiobjective Optimization; Abraham, A., Jain, L., Goldberg, R., Eds.; Advanced Information and Knowledge Processing; Springer: London, UK, 2005; pp. 105–145. ISBN 978-1-85233-787-2. [Google Scholar]
  152. Deb, K.; Jain, H. An Evolutionary Many-Objective Optimization Algorithm Using Reference-Point-Based Nondominated Sorting Approach, Part I: Solving Problems with Box Constraints. IEEE Trans. Evol. Comput. 2014, 18, 577–601. [Google Scholar] [CrossRef]
  153. Emmerich, M.T.M.; Deutz, A.H. A Tutorial on Multiobjective Optimization: Fundamentals and Evolutionary Methods. Nat. Comput. 2018, 17, 585–609. [Google Scholar] [CrossRef] [PubMed]
  154. Mendoza, F.; Bernal-Agustin, J.L.; Dominguez-Navarro, J.A. NSGA and SPEA Applied to Multiobjective Design of Power Distribution Systems. IEEE Trans. Power Syst. 2006, 21, 1938–1945. [Google Scholar] [CrossRef]
  155. Zitzler, E.; Laumanns, M.; Thiele, L. SPEA2: Improving the Strength Pareto Evolutionary Algorithm; ETH Zurich: Zürich, Switzerland, 2001; p. 21. [Google Scholar]
  156. Liu, S.; Tao, R.; Tam, C.M. Optimizing Cost and CO2 Emission for Construction Projects Using Particle Swarm Optimization. Habitat Int. 2013, 37, 155–162. [Google Scholar] [CrossRef]
  157. Gan, V.J.L.; Lo, I.M.C.; Ma, J.; Tse, K.T.; Cheng, J.C.P.; Chan, C.M. Simulation Optimisation towards Energy Efficient Green Buildings: Current Status and Future Trends. J. Clean. Prod. 2020, 254, 120012. [Google Scholar] [CrossRef]
  158. Raschka, S.; Patterson, J.; Nolet, C. Machine Learning in Python: Main Developments and Technology Trends in Data Science, Machine Learning, and Artificial Intelligence. Information 2020, 11, 193. [Google Scholar] [CrossRef]
  159. MATLAB. Available online: https://www.mathworks.com/products/matlab.html (accessed on 7 April 2024).
  160. Start [Jeplus.Org]. Available online: http://www.jeplus.org/wiki/doku.php?id=start (accessed on 7 April 2024).
  161. Multiobjective Optimization-MATLAB & Simulink. Available online: https://www.mathworks.com/help/gads/multiobjective-optimization.html (accessed on 7 April 2024).
  162. Cui, Y.; Geng, Z.; Zhu, Q.; Han, Y. Review: Multi-Objective Optimization Methods and Application in Energy Saving. Energy 2017, 125, 681–704. [Google Scholar] [CrossRef]
  163. Veldhuizen, D.A.V.; Lamont, G.B. Multiobjective Evolutionary Algorithm Research: A History and Analysis; TR-98e03; Air Force Institute of Technology: Dayton, OH, USA, 1998. [Google Scholar]
  164. Zitzler, E.; Thiele, L.; Laumanns, M.; Fonseca, C.M.; Da Fonseca, V.G. Performance Assessment of Multiobjective Optimizers: An Analysis and Review. IEEE Trans. Evol. Comput. 2003, 7, 117–132. [Google Scholar] [CrossRef]
  165. Pinzon Amorocho, J.A.; Hartmann, T. A Multi-Criteria Decision-Making Framework for Residential Building Renovation Using Pairwise Comparison and TOPSIS Methods. J. Build. Eng. 2022, 53, 104596. [Google Scholar] [CrossRef]
  166. Chauhan, A.; Vaish, R. A Comparative Study on Decision Making Methods with Interval Data. J. Comput. Eng. 2014, 2014, 93074. [Google Scholar] [CrossRef]
  167. Rokbani, N.; Neji, B.; Slim, M.; Mirjalili, S.; Ghandour, R. A Multi-Objective Modified PSO for Inverse Kinematics of a 5-DOF Robotic Arm. Appl. Sci. 2022, 12, 7091. [Google Scholar] [CrossRef]
  168. Francart, N.; Malmqvist, T.; Hagbert, P. Climate Target Fulfilment in Scenarios for a Sustainable Swedish Built Environment beyond Growth. Futures 2018, 98, 1–18. [Google Scholar] [CrossRef]
  169. Ahmad, M.W.; Mourshed, M.; Rezgui, Y. Trees vs Neurons: Comparison between Random Forest and ANN for High-Resolution Prediction of Building Energy Consumption. Energy Build. 2017, 147, 77–89. [Google Scholar] [CrossRef]
  170. Wang, Z.; Liu, J.; Zhang, Y.; Yuan, H.; Zhang, R.; Srinivasan, R.S. Practical Issues in Implementing Machine-Learning Models for Building Energy Efficiency: Moving beyond Obstacles. Renew. Sustain. Energy Rev. 2021, 143, 110929. [Google Scholar] [CrossRef]
  171. Singaravel, S.; Suykens, J.; Geyer, P. Deep-Learning Neural-Network Architectures and Methods: Using Component-Based Models in Building-Design Energy Prediction. Adv. Eng. Inform. 2018, 38, 81–90. [Google Scholar] [CrossRef]
  172. Khalil, M.; McGough, A.S.; Pourmirza, Z.; Pazhoohesh, M.; Walker, S. Machine Learning, Deep Learning and Statistical Analysis for Forecasting Building Energy Consumption—A Systematic Review. Eng. Appl. Artif. Intell. 2022, 115, 105287. [Google Scholar] [CrossRef]
  173. Qian, F.; Gao, W.; Yang, Y.; Yu, D. Potential Analysis of the Transfer Learning Model in Short and Medium-Term Forecasting of Building HVAC Energy Consumption. Energy 2020, 193, 116724. [Google Scholar] [CrossRef]
  174. Lu, C.; Li, S.; Lu, Z. Building Energy Prediction Using Artificial Neural Networks: A Literature Survey. Energy Build. 2022, 262, 111718. [Google Scholar] [CrossRef]
  175. Man, Q.; Yu, H.; Feng, K.; Olofsson, T.; Lu, W. Transfer of Building Retrofitting Evaluations for Data-Scarce Conditions: An Empirical Study for Sweden to China. Energy Build. 2024, 310, 114041. [Google Scholar] [CrossRef]
  176. Fahlstedt, O.; Temeljotov-Salaj, A.; Lohne, J.; Bohne, R.A. Holistic Assessment of Carbon Abatement Strategies in Building Refurbishment Literature—A Scoping Review. Renew. Sustain. Energy Rev. 2022, 167, 112636. [Google Scholar] [CrossRef]
  177. Bian, Y.; Chen, Y.; Sun, Y.; Ma, Y.; Yu, D.; Leng, T. Simulation of Daylight Availability, Visual Comfort and View Clarity for a Novel Window System with Switchable Blinds in Classrooms. Build. Environ. 2023, 235, 110243. [Google Scholar] [CrossRef]
  178. Xu, S.; Li, C.; He, W.; Chu, W.; Hu, Z.; Lu, B. Experimental Study of Bifacial Photovoltaic Wall System Incorporating Thermochromic Material. Sustain. Cities Soc. 2024, 106, 105372. [Google Scholar] [CrossRef]
  179. Bui, D.-K.; Nguyen, T.N.; Ghazlan, A.; Ngo, T.D. Biomimetic Adaptive Electrochromic Windows for Enhancing Building Energy Efficiency. Appl. Energy 2021, 300, 117341. [Google Scholar] [CrossRef]
  180. Cheng, Y.; Chen, X.; Xu, B.; Pei, G.; Jiao, D. Experimental Analysis of Building Envelope Integrating Phase Change Material and Cool Paint under a Real Environment in Autumn. J. Clean. Prod. 2024, 461, 142674. [Google Scholar] [CrossRef]
  181. Quan, S.J.; Wu, J.; Wang, Y.; Shi, Z.; Yang, T.; Yang, P.P.-J. Urban Form and Building Energy Performance in Shanghai Neighborhoods. Energy Procedia 2016, 88, 126–132. [Google Scholar] [CrossRef]
  182. Ekici, B.; Cubukcuoglu, C.; Turrin, M.; Sariyildiz, I.S. Performative Computational Architecture Using Swarm and Evolutionary Optimisation: A Review. Build. Environ. 2019, 147, 356–371. [Google Scholar] [CrossRef]
  183. Octopus. Available online: https://www.grasshopper3d.com/groups/group/show?groupUrl=octopus (accessed on 7 April 2024).
  184. Fan, Z.; Liu, M.; Tang, S. A Multi-Objective Optimization Design Method for Gymnasium Facade Shading Ratio Integrating Energy Load and Daylight Comfort. Build. Environ. 2022, 207, 108527. [Google Scholar] [CrossRef]
  185. Nasrollahzadeh, N. Comprehensive Building Envelope Optimization: Improving Energy, Daylight, and Thermal Comfort Performance of the Dwelling Unit. J. Build. Eng. 2021, 44, 103418. [Google Scholar] [CrossRef]
  186. Ji, Y.; Wang, W.; He, Y.; Li, L.; Zhang, H.; Zhang, T. Performance in Generation: An Automatic Generalizable Generative-Design-Based Performance Optimization Framework for Sustainable Building Design. Energy Build. 2023, 298, 113512. [Google Scholar] [CrossRef]
  187. Chen, R.; Samuelson, H.; Zou, Y.; Zheng, X.; Cao, Y. Improving Building Resilience in the Face of Future Climate Uncertainty: A Comprehensive Framework for Enhancing Building Life Cycle Performance. Energy Build. 2024, 302, 113761. [Google Scholar] [CrossRef]
  188. Madushika, U.G.D.; Lu, W. Green Retrofitting Application in Developing Economies: State of the Art and Future Research Directions. Energy Build. 2023, 301, 113712. [Google Scholar] [CrossRef]
  189. Tian, Z.; Zhang, X.; Wei, S.; Du, S.; Shi, X. A Review of Data-Driven Building Performance Analysis and Design on Big on-Site Building Performance Data. J. Build. Eng. 2021, 41, 102706. [Google Scholar] [CrossRef]
  190. Singh, M.K.; Attia, S.; Mahapatra, S.; Teller, J. Assessment of Thermal Comfort in Existing Pre-1945 Residential Building Stock. Energy 2016, 98, 122–134. [Google Scholar] [CrossRef]
  191. Alonso, A.; Patricio, J.; Suárez, R.; Escandón, R. Acoustical Retrofit of Existing Residential Buildings: Requirements and Recommendations for Sound Insulation between Dwellings in Europe and Other Countries Worldwide. Build. Environ. 2020, 174, 106771. [Google Scholar] [CrossRef]
  192. Aftab, M.; Chen, C.; Chau, C.-K.; Rahwan, T. Automatic HVAC Control with Real-Time Occupancy Recognition and Simulation-Guided Model Predictive Control in Low-Cost Embedded System. Energy Build. 2017, 154, 141–156. [Google Scholar] [CrossRef]
  193. Hong, Y.; Ezeh, C.I.; Deng, W.; Hong, S.-H.; Peng, Z. Building Energy Retrofit Measures in Hot-Summer–Cold-Winter Climates: A Case Study in Shanghai. Energies 2019, 12, 3393. [Google Scholar] [CrossRef]
  194. Li, L.; Ma, Q.; Gao, W.; Wei, X. Incorporating Users’ Adaptive Behaviors into Multi-Objective Optimization of Shading Devices: A Case Study of an Office Room in Qingdao. Energy Build. 2023, 301, 113683. [Google Scholar] [CrossRef]
  195. De Wilde, P.; Coley, D. The Implications of a Changing Climate for Buildings. Build. Environ. 2012, 55, 1–7. [Google Scholar] [CrossRef]
  196. Daly, D.; Cooper, P.; Ma, Z. Implications of Global Warming for Commercial Building Retrofitting in Australian Cities. Build. Environ. 2014, 74, 86–95. [Google Scholar] [CrossRef]
  197. Balezentis, T.; Siksnelyte-Butkiene, I.; Streimikiene, D. Stakeholder Involvement for Sustainable Energy Development Based on Uncertain Group Decision Making: Prioritizing the Renewable Energy Heating Technologies and the BWM-WASPAS-IN Approach. Sustain. Cities Soc. 2021, 73, 103114. [Google Scholar] [CrossRef]
  198. Beccali, M.; Cellura, M.; Fontana, M.; Longo, S.; Mistretta, M. Energy Retrofit of a Single-Family House: Life Cycle Net Energy Saving and Environmental Benefits. Renew. Sustain. Energy Rev. 2013, 27, 283–293. [Google Scholar] [CrossRef]
  199. Allegrini, J.; Dorer, V.; Carmeliet, J. Influence of the Urban Microclimate in Street Canyons on the Energy Demand for Space Cooling and Heating of Buildings. Energy Build. 2012, 55, 823–832. [Google Scholar] [CrossRef]
  200. Santamouris, M.; Papanikolaou, N.; Livada, I.; Koronakis, I.; Georgakis, C.; Argiriou, A.; Assimakopoulos, D.N. On the Impact of Urban Climate on the Energy Consumption of Buildings. Sol. Energy 2001, 70, 201–216. [Google Scholar] [CrossRef]
  201. Reitberger, R.; Palm, N.; Palm, H.; Lang, W. Urban Systems Exploration: A Generic Process for Multi-Objective Urban Planning to Support Decision Making in Early Design Phases. Build. Environ. 2024, 254, 111360. [Google Scholar] [CrossRef]
Figure 1. Literature search and screening process.
Figure 1. Literature search and screening process.
Energies 17 04641 g001
Figure 2. Quantitative trends and geographical distribution of publications.
Figure 2. Quantitative trends and geographical distribution of publications.
Energies 17 04641 g002
Figure 3. Building types and data-driven methods of research. (a) Building types; (b) data-driven methods.
Figure 3. Building types and data-driven methods of research. (a) Building types; (b) data-driven methods.
Energies 17 04641 g003
Figure 4. Top 9 keywords with the strongest citation bursts (the red-line segments indicate periods of significant citation of keywords).
Figure 4. Top 9 keywords with the strongest citation bursts (the red-line segments indicate periods of significant citation of keywords).
Energies 17 04641 g004
Figure 5. General process of building retrofit and performance optimizations (the dash lines represent the distinction between each step and the dash arrows indicate the optional processes based on computational requirements).
Figure 5. General process of building retrofit and performance optimizations (the dash lines represent the distinction between each step and the dash arrows indicate the optional processes based on computational requirements).
Energies 17 04641 g005
Figure 6. Heatmap distribution of CE objectives with other research goals.
Figure 6. Heatmap distribution of CE objectives with other research goals.
Energies 17 04641 g006
Figure 7. Frequency ranking of CE objectives with other research goals.
Figure 7. Frequency ranking of CE objectives with other research goals.
Energies 17 04641 g007
Figure 8. Retrofit variables in studies.
Figure 8. Retrofit variables in studies.
Energies 17 04641 g008
Figure 9. Quantitative distribution of BPO methods.
Figure 9. Quantitative distribution of BPO methods.
Energies 17 04641 g009
Figure 10. Correspondence among simulation tools, optimization algorithms, and tools.
Figure 10. Correspondence among simulation tools, optimization algorithms, and tools.
Energies 17 04641 g010
Figure 11. Typical examples of the Pareto front [161]. (a) Two-dimensional Pareto front; (b) three-dimensional Pareto front (the blue circles represent the solutions to the MOO problem).
Figure 11. Typical examples of the Pareto front [161]. (a) Two-dimensional Pareto front; (b) three-dimensional Pareto front (the blue circles represent the solutions to the MOO problem).
Energies 17 04641 g011
Figure 12. Examples of TOPSIS and Utopia-point methods in 2D coordinates. (a) TOPSIS method [166]; (b) Utopia-point method [167] (the purple circles represent the solutions to the MOO problem).
Figure 12. Examples of TOPSIS and Utopia-point methods in 2D coordinates. (a) TOPSIS method [166]; (b) Utopia-point method [167] (the purple circles represent the solutions to the MOO problem).
Energies 17 04641 g012
Table 1. Literature search keywords.
Table 1. Literature search keywords.
TermKeywords
Term 1“building renovation” OR “building reconstruction” OR “building retrofit *” OR “building refurbishment” OR “building repairment” OR “building restoration” OR “building upgrade” OR “building renewal” OR “building improvement” OR “building reformation”
Term 2multi-objective OR multi-criteria OR optimization
Term 3“carbon emission” OR “carbon mitigation” OR “CO2 emission” OR “CO2 mitigation” OR “greenhouse gas” OR “global warming” OR “environmental impact” OR “sustainable development”
Note: The wildcard “*” indicates a fuzzy search and is used to replace words with multiple spelling variants, such as “retrofit” and “retrofitting”.
Table 2. Summary of previous research studies (optimization objectives and variables).
Table 2. Summary of previous research studies (optimization objectives and variables).
Refs.LocationBuilding TypeOptimization ObjectiveOptimization Variable
[23]UKOffice buildingLCCs, LCEs, and
LCCEs
Insulation material area of roof and exterior wall,
equipment and energy system,
PV panel area, and solar thermal device.
[24]SwitzerlandResidential buildingCEs and ACsU-/R-value of roof, exterior wall, and ground; window type;
equipment and energy system; PV system; and solar thermal device.
[25]ItalyOffice buildingECs, TDHs, GCs, and GHGEsSurface material characteristics of roof and exterior wall,
insulation material thickness of roof and exterior wall,
window type, equipment and energy system,
sunshade component, and PV panel angle and area.
[38]UKOffice buildingLCCEs and OCEInsulation material type of roof and exterior wall,
equipment and energy system, and solar thermal device.
[39]UKOffice buildingLCCs, LCEs, and LCCEsInsulation material type and area of roof and exterior wall,
window-to-wall ratio, equipment and energy system,
PV panel area, and solar thermal device.
[40]ChinaShopping mallOCEsU/R-value of exterior wall, Glass material,
Sunshade component and equipment and energy system.
[41]IranResidential buildingCEs and TDHsInsulation material thickness of roof and exterior wall,
insulation material thickness and type of ground,
window type, airtightness, and equipment and energy system.
[42]CanadaOffice buildingECs and CEsInsulation material type of roof, exterior wall, and floor;
window type; airtightness; and equipment and energy system.
[43]FinlandOffice buildingLCCs, RCs, CEs, and TDHsInsulation material thickness of roof and exterior wall,
window type, sunshade component,
equipment and energy system, and PV system.
[44]IranResidential buildingECs and the GWPInsulation material type and thickness of exterior wall and
exterior wall type (combination of different materials).
[45]CanadaEducational buildingECs, LCCs, and LCAsType of roof and exterior wall, glass material, airtightness,
window opening percentage, and equipment and energy system.
[46]KoreaResidential buildingRCs, LCCs, LCCEs, and CERsInsulation material type and thickness of exterior wall,
window type, sunshade component,
and equipment and energy system.
[47]FranceEducational buildingECs, TDHs, RCs, and CEsType of roof, floor, ground, and interior and exterior wall;
window type, and sunshade component.
[48]EuropeResidential buildingECs, RCs, OCs, and CEsSurface material characteristics of roof and exterior wall,
window type, sunshade component, sunspace,
building form, PV panel angle and area, and solar thermal device.
[49]FinlandResidential buildingECs, LCCs, and CEsInsulation material thickness of roof and exterior wall,
window type, door material, PV panel area,
solar thermal device, and equipment and energy system.
[50]IranResidential buildingCEs, WCs, LCCs, and TDHsInsulation material type and thickness of roof and exterior wall,
glass material, filling gas, PV panel area,
and equipment and energy system.
[51]ChinaResidential buildingCEs, TDHs, and GCsSurface material characteristics, insulation material type and thickness of roof and exterior wall, window type,
sunshade component, sunspace, and PV panel angle and area.
[52]ChinaResidential buildingECs, RCs, and CERsInsulation material type of roof, exterior wall and floor,
glass material, window-to-wall ratio, and sunspace.
[53]EstoniaResidential buildingGCs, ECs, and LCCEsInsulation material thickness of exterior wall,
surface material characteristics of roof,
window type, door material, and building form.
[54]KoreaEducational buildingECs, CEs, RCs, and TDHsType of roof, floor, ground, ceiling, and interior and exterior wall;
window type; and equipment and energy system.
[55]ChinaResidential buildingThe GWP, LCCs, and TDHsInsulation material type and thickness of roof and exterior wall,
window type, window-to-wall ratio, and sunshade component.
[56]ChinaResidential buildingECs, LCCEs, and LCCsInsulation material type and thickness of floor and exterior wall,
glass material, window-to-wall ratio,
sunshade component, and Airtightness.
[57]UKResidential buildingLCCEs and LCCsInsulation material thickness, exterior wall type,
and window-to-wall ratio.
[58]SwedenResidential buildingLCEs, LCCEs, and LCCsInsulation material type and thickness of exterior wall, roof, and ground and window type.
[59]SwitzerlandResidential buildingACs CEsU-/R-value of roof, floor, and exterior wall; window type;
PV panel area; and solar thermal device,
and equipment and energy system.
[60]CanadaResidential buildingLCCEs and LCCsInsulation material type of ceiling and exterior wall,
window frame material, door material,
airtightness, and equipment and energy system.
[61]UKNon-domestic buildingBERType of roof and exterior wall, window type,
and equipment and energy system.
[62]ItalyResidential buildingECs, OCs, RCs, and CEsInsulation material thickness of roof, floor, and exterior wall;
surface material characteristics of roof and exterior wall;
PV panel angle and area; glass material; sunshade component;
building form; sunspace; and solar thermal device.
[63]IranResidential buildingCEs and TDHsInsulation material thickness of roof, ground, and exterior wall;
window type, airtightness, and equipment and energy system.
[64]DenmarkResidential buildingECs, the GWP, OCs, and RCsInsulation material type and thickness of interior wall,
insulation material type and thickness of roof and exterior wall,
surface material characteristics of roof and exterior wall,
window frame material, glass material, PV panel area,
solar thermal device, and equipment and energy system.
[65]ItalyResidential buildingRCs, OCs, ECs, and CEsInsulation material thickness of roof, floor, and exterior wall;
surface material characteristics of exterior wall;
PV panel angle and area; sunshade component,
building form, sunspace, and solar thermal device.
[66]Bosnia and HerzegovinaResidential buildingECs, CEs, and RCsInsulation material thickness of ceiling and exterior wall,
window type, and equipment and energy system.
[67]ChinaOffice buildingECs, CEs, and TDHsPV panel angle and area and equipment and energy system.
[68]SwitzerlandResidential buildingLCCs and GHGEsType of roof and exterior wall, window type, airtightness,
PV system, solar thermal device, and equipment and energy system.
[69]GermanyResidential buildingACs and CEsType of roof and exterior wall, window type, PV system,
solar thermal device, and equipment and energy system.
[70]ChinaOffice buildingECs, CEs, and OCsU-/R-value of roof and exterior wall, window type,
and equipment and energy system.
[71]GreeceResidential buildingGHGEs and LCCsInsulation material thickness of roof, ground, and exterior wall;
window type, PV system, solar thermal device,
and equipment and energy system.
[72]ChinaOffice buildingLCCEsInsulation material thickness of roof and exterior wall,
surface material characteristics of exterior wall, window type,
PV panel area, and equipment and energy system.
[73]ChinaEducational buildingECs and LCCEsType of roof, floor, and exterior wall; filling gas; building form;
insulation material thickness of floor and exterior wall;
window frame material; glass material; building form;
insulation material thickness of roof; window-to-wall ratio; sunshade component; PV panel area;
and equipment and energy system.
[74]USAResidential buildingGHGEs, WCs, the SQOL, and LCCsU-/R-value of roof, ceiling and exterior wall, glass material,
window-to-wall ratio, and equipment and energy system.
[75]UKResidential buildingLCCEs and LCCsType of roof, floor, ceiling, and interior and exterior wall and
window type.
[76]CanadaEducational buildingECs, LCCs, and LCAsType of roof and exterior wall, glass material,
window frame material, window-to-wall ratio, airtightness,
window opening percentage, and equipment and energy system.
[77]CanadaOffice buildingECs, ECEs, and LCCsType of roof and exterior wall, glass material,
window frame material, window-to-wall ratio, airtightness,
sunshade component, and equipment and energy system.
[78]ChinaOffice buildingECs, CEs, and LCCsInsulation material type and thickness of roof and exterior wall and window type.
[79]SwitzerlandResidential buildingLCCs and LCAsInsulation material type of ceiling and exterior wall;
insulation material thickness of ceiling, floor, and exterior wall;
glass material; and window frame material.
Table 3. Summary of previous research studies (machine learning and sensitivity analysis method).
Table 3. Summary of previous research studies (machine learning and sensitivity analysis method).
Refs.LocationBuilding TypeMachine Learning Method (Accuracy)Sensitivity Analysis Method
[23]UKOffice building--
[24]SwitzerlandResidential buildingANN
(R2 = 0.94)
-
[25]ItalyOffice building--
[38]UKOffice building--
[39]UKOffice building-LSA
[40]ChinaShopping mall-LSA
[41]IranResidential building-GSA (DOE)
[42]CanadaOffice building-LSA
[43]FinlandOffice building--
[44]IranResidential building--
[45]CanadaEducational buildingANN (MSE1 = 0.016 and
MSE2 = 0.056)
-
[46]KoreaResidential building--
[47]FranceEducational building--
[48]EuropeResidential building--
[49]FinlandResidential building--
[50]IranResidential building--
[51]ChinaResidential building-GSA
(PCC and SRRC)
[52]ChinaResidential building--
[53]EstoniaResidential building--
[54]KoreaEducational building--
[55]ChinaResidential buildingDNN (R2 > 0.99,
CV (RMSE) ≤ 1%, and
NMBE ≤ 0.2%)
GSA
[56]ChinaResidential building--
[57]UKResidential building--
[58]SwedenResidential building--
[59]SwitzerlandResidential building--
[60]CanadaResidential building--
[61]UKNon-domestic buildingGBRT
(RMSE = 1.7%)
LSA
[62]ItalyResidential building-GSA (SRRC)
[63]IranResidential building-GSA (DOE)
[64]DenmarkResidential building--
[65]ItalyResidential building-GSA (SRRC)
[66]Bosnia and HerzegovinaResidential building--
[67]ChinaOffice building-GSA (SRC)
[68]SwitzerlandResidential building--
[69]GermanyResidential building--
[70]ChinaOffice building-GSA (Morris)
[71]GreeceResidential building--
[72]ChinaOffice building--
[73]ChinaEducational buildingANN
(MRE = 1.57%
R2 = 0.94)
-
[74]USAResidential building--
[75]UKResidential building--
[76]CanadaEducational building--
[77]CanadaOffice buildingMVLR and MARSs
(MAPE = 0.2–1.8%)
-
[78]ChinaOffice building--
[79]SwitzerlandResidential buildingGaussian process modelling (Kriging)GSA (Sobol)
Table 4. Summary of previous research studies (optimization and decision-making method).
Table 4. Summary of previous research studies (optimization and decision-making method).
Refs.LocationBuilding TypeOptimization MethodDecision-Making Method
[23]UKOffice buildingPSO-
[24]SwitzerlandResidential buildingMILP-
[25]ItalyOffice buildingNSGA-II-
[38]UKOffice buildingPSO-
[39]UKOffice buildingPSO-
[40]ChinaShopping mallRegression-
[41]IranResidential buildingNSGA-II-
[42]CanadaOffice building--
[43]FinlandOffice buildingPareto-Archive and NSGA-II-
[44]IranResidential buildingFitness Comparison-
[45]CanadaEducational buildingNSGA-II-
[46]KoreaResidential buildingiMOO score-
[47]FranceEducational buildingNSGA-II-
[48]EuropeResidential buildingaNSGA-II and
pNSGA-II
Utopia point
[49]FinlandResidential buildingPareto-Archive and NSGA-II-
[50]IranResidential buildingprNSGA-IIITOPSIS
[51]ChinaResidential buildingSPEA2Utopia point
[52]ChinaResidential building-Entropy method
(Weight of CERs is 30.95%)
[53]EstoniaResidential buildingRegression-
[54]KoreaEducational buildingNSGA-II/III and
MOEA/D
-
[55]ChinaResidential buildingNSGA-IITOPSIS
(Weight of the GWP is 37.29%)
[56]ChinaResidential buildingNSGA-II-
[57]UKResidential buildingNSGA-II-
[58]SwedenResidential buildingNSGA-II-
[59]SwitzerlandResidential buildingGA and MILP-
[60]CanadaResidential buildingNSGA-
[61]UKNon-domestic buildingGA-
[62]ItalyResidential buildingaNSGA-IIUtopia point
[63]IranResidential buildingEWSOA-
[64]DenmarkResidential buildingOmni-OptimizerUtopia point
[65]ItalyResidential buildingaNSGA-IIUtopia point
[66]Bosnia and HerzegovinaResidential buildingNSGA-IIIDesirability function
(Weight of CEs is 30%)
[67]ChinaOffice buildingNSGA-II-
[68]SwitzerlandResidential buildingϵ-constraint-
[69]GermanyResidential buildingϵ-constraint-
[70]ChinaOffice building--
[71]GreeceResidential buildingMOGA-
[72]ChinaOffice buildingNOP and MILP-
[73]ChinaEducational buildingSEGA-
[74]USAResidential buildingGA-
[75]UKResidential buildingNSGA-II-
[76]CanadaEducational buildingNSGA-II-
[77]CanadaOffice building--
[78]ChinaOffice buildingAHP-
[79]SwitzerlandResidential buildingNSGA-II-
Table 5. The use of BPS tools in the literature.
Table 5. The use of BPS tools in the literature.
Simulation ToolReferences
Designbuilder[25,40,41,44,45,46,50,63,66,68,73,76,77,78]
TRNSYS[39,47,67,71]
SIMEB[42]
IDA ICE[43,49,53]
SketchUp—OpenStudio[48,52,57,62,65,70]
Grasshopper—Honeybee[51,55,56,58,72]
EnergyPlus[54,59,74,75]
HOT2000[60]
Table 6. Commonly used error-evaluation indicators.
Table 6. Commonly used error-evaluation indicators.
Evaluation IndicatorsGuidelineMonthly CriteriaHourly CriteriaReferences
MBEASHRAE±5%±10%[46,51,55,56,67]
IPMVP-±5%
FEMP±5%±10%
CV (RMSE)ASHRAE15%30%[46,50,51,55,56,67,77]
IPMVP-20%
FEMP15%30%
Table 7. Classification and application of sensitivity analysis methods.
Table 7. Classification and application of sensitivity analysis methods.
Sensitivity Analysis MethodReferences
LSA [39,40,42,61]
GSAMetamodel-based method[41,63]
Regression-based method[51,62,65,67]
Variance-based method[55,79]
Density-based method[55]
Screening-based method[70]
Table 8. The use of optimization tools in the literature.
Table 8. The use of optimization tools in the literature.
Optimization ToolReferencesOptimization ToolReferences
MATLAB[25,50,72,74]Python[48,56,58,60,62,65,71,73,77]
SPSS[40]Octopus[51]
jEPlus + EA[41,57,67]JESS + JEA[63]
MOBO[43,49,64]CPLEX[68]
Excel-VBA[46]Gurobi[69]
MultiOpt[47]PLOOTO[75]
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

Share and Cite

MDPI and ACS Style

Luo, S.-L.; Shi, X.; Yang, F. A Review of Data-Driven Methods in Building Retrofit and Performance Optimization: From the Perspective of Carbon Emission Reductions. Energies 2024, 17, 4641. https://doi.org/10.3390/en17184641

AMA Style

Luo S-L, Shi X, Yang F. A Review of Data-Driven Methods in Building Retrofit and Performance Optimization: From the Perspective of Carbon Emission Reductions. Energies. 2024; 17(18):4641. https://doi.org/10.3390/en17184641

Chicago/Turabian Style

Luo, Shu-Long, Xing Shi, and Feng Yang. 2024. "A Review of Data-Driven Methods in Building Retrofit and Performance Optimization: From the Perspective of Carbon Emission Reductions" Energies 17, no. 18: 4641. https://doi.org/10.3390/en17184641

APA Style

Luo, S. -L., Shi, X., & Yang, F. (2024). A Review of Data-Driven Methods in Building Retrofit and Performance Optimization: From the Perspective of Carbon Emission Reductions. Energies, 17(18), 4641. https://doi.org/10.3390/en17184641

Note that from the first issue of 2016, this journal uses article numbers instead of page numbers. See further details here.

Article Metrics

Back to TopTop