The Impact of Architects’ Reasoning on Early Design Decision-Making for Energy-Efficient Buildings
1
College of Built Environment, Birmingham City University, Birmingham B4 7AP, UK
2
School of Architecture Building and Civil Engineering, Loughborough University, Leicestershire LE11 3TU, UK
*
Author to whom correspondence should be addressed.
Sustainability 2024, 16(18), 8220; https://doi.org/10.3390/su16188220
Submission received: 13 August 2024
/
Revised: 11 September 2024
/
Accepted: 16 September 2024
/
Published: 21 September 2024
(This article belongs to the Special Issue Novel Technologies and Digital Design in Smart Construction)
Abstract
:Architects arguably have the greatest influence on the design of buildings. One of the key factors that make it hard to improve the energy efficiency of buildings is the use of architects’ reasoning by architects at the early design stage. There is a need to assess the impact of architects’ reasoning on the energy performance of the designed building. To this end, this research was conducted in two phases. Firstly, the most influential design parameters, in terms of energy efficiency, were identified and used to develop a design exercise issued to a sample of practising architects in the north of Algeria. Design exercise participants were required to minimise expected energy consumption along with the construction cost. Secondly, computer-generated dynamic design optimisation for the same design task was conducted in DesignBuilder v6. 1 .8. The computer-generated designs decisively outperformed the human-generated designs. The experienced architects achieved the least-performing designs rather than those with less experience.
1. Introduction
To design energy-efficient buildings that respond to the current challenges of sustainability, particular aspects have to be considered from the early stage of design. Important decisions are made early in the design that significantly affect a building’s energy performance over its lifecycle [1,2,3]. These decisions are in the form of several design variables that that need to be fixed early in the design process. By considering energy efficiency at an early stage, the architect has the greatest influence over the lifecycle performance of a building. Influential energy-efficient design variables such as building orientation and building fabric are usually decided early on. Once these decisions have been made, it is difficult to make subsequent revisions in order to enhance energy efficiency. Although energy performance is of paramount importance, it cannot be considered in isolation without regard for other design priorities.
As energy aspirations become increasingly ambitious, the intricacy of these aspirations and the complexity of building design challenge the architect’s built-up experience [4]. Ercan and Elias-Ozkan [5] argue that the integration of energy performance measures needs to start from the conceptual design stage. The architect’s reasoning is therefore crucially important during the early design stage to ensure energy-informed decisions are made in a timely manner.
The energy performance gap between design intent and actual performance during operation originates in early design decision-making, arguably due to the cognitive process employed by architects’ minds while working on design problems [6]. Commonly, the early design process is controlled by architects alone, usually without prior energy-efficiency expertise [7]. This was supported by Rezaee et al. [8], who reported the wide confidence intervals for performance predictions in thermal building performance during early design stages; they reported confidence levels well below 50%. Lacking expertise in building physics, architects’ choices are predominantly guided by intuition and experience that fails to account for the implications of design decisions on building energy performance. Consequently, architects tend to favour conventional solutions (known by practice) even though their energy performance is not optimal. The amount of information humans may consider at a time is limited to a few pieces of information, which can obstruct human design problem-solving.
Such limitations have been observed as humans solve increasingly complex parametric design problems. Hirschi and Frey [9] report experiments showing that human problem-solving performance declines considerably as the number of considered variables increases. Much research has been conducted into the architectural design process and the methods designers use to solve design problems [10,11,12]. Such work signposts the scope for significant original research to test architects’ reasoning concerning energy performance during the early design stage. By reasoning, the knowledge repertoire that is used by architects to make early design decisions is meant to inform and guide their initial creative process.
Within a background of more stringent building energy targets, designers must approach with care the task of designing their buildings. This research aims to assess the tacit reasoning of architects and its role in achieving energy-efficient design. Therefore, the primary building design performance objective considered in this paper is energy consumption. Cost was added as a second objective function, because, in real-world projects, it is hard to imagine a scenario in which the designer is only concerned with energy performance but unconcerned with the cost of the resulting design. The most important design variables that have the highest impact on energy performance (as identified from an online questionnaire survey) were then used to design an exercise activity whereby each participant was given the base case design model and asked to modify it to achieve the two competing objectives.
2. Architects’ Reasoning in Design Decision-Making
Much of the knowledge in people’s heads is constructed within the system in which they operate, and, therefore, this reasoning mechanism is the property of the system [13]. In the case of building performance, energy simulation and optimisation are typically employed to create design solutions that meet the desired design objectives. As a result, the majority of the simulation software that is generally used supports optimisation tasks that are effective only late in the design process. Energy performance simulation generally requires precise building geometry as well as data about the fabric characteristics of building elements. The early design stage, however, is characterised by high-level activity with regards to the architectural aspects of the design, where such design precision is not yet possible. Petersen [14] suggests concentrating significant effort in the early design stage by handling the design variables that architects can access and control at this early stage. However, the lack of building physics expertise has always hindered design decision-makers [15]. Without being able to investigate design alternatives against performance targets early on in the design process, the resulting design may not be optimal due to the multiple trade-offs to be made. In the absence of precise simulation predictions, architects tend to use their reasoning based on experience to develop a design alternative. Thus, in this research, architects’ reasoning is singled out as the factor which differentiates design decision-making by humans versus the precise optimisation of computers. DesignBuilder was used as the optimisation platform, as it has the capability of running optimisation with multiple design objectives, for the following research: energy consumption and construction cost.
3. Method
Our research method adopts a descriptive design models paradigm, which emphasises the process of generating a solution from an early stage. This paradigm emerged as a response to the notion that what many designers do in actual practice does not completely algin with the classical systematic pattern of design activities. De Wilde [6] observes that descriptive decision-making does not always ensure optimum energy performance-based design. Therefore, this research assesses the performance of this naturalistic decision-making driven by architects’ reasoning. Two research methods were combined in an attempt to isolate the effect of tacit reasoning by human architects. Firstly, a design exercise experiment was conducted to trigger naturalistic design decision-making in practising architects. Secondly, computational dynamic simulation and optimisation was applied to the same design problem to explore the entire design space thoroughly. Figure 1 outlines the research design.
3.1. Human Architects Design Experiment
Researchers use several approaches for empirically studying human designers, including verbal protocols, case studies, and design experiments [11]. Design experiments were adopted, because they enable the precise study of specific design processes with rigorous statistical comparisons, rather than case studies and verbal protocols that may contain more conflating contextual variables that obscure the validity of conclusions. Figure 2 presents the experimental procedure for the design experiments.
In the experiment, practicing architects were asked to complete a design exercise, balancing the two competing objectives of minimising energy consumption and reducing construction cost. Each architect’s years of experience was measured as an independent variable in the design experiment. Experimental comparisons of novices and experts are common in design research [16,17].
A design task base case model of a residential building block was created. Designers were asked to modify the base case design considering two design objectives: to minimise energy consumption and to minimise construction cost, with a particular emphasis placed on an energy-efficient design. Measures of performance (MOP) indicators were used to evaluate architects’ reasoning in terms of building energy and cost performance.
3.2. Base Case Model Definition
3.2.1. Weather Profile
Weather conditions directly impact the results of any energy performance simulation [18,19]. The location of this study was chosen to be Algeria. Direct access to the Algerian National Council of Architects by the first author facilitated the recruitment of research participants for the design exercise. The location was also chosen due to the importance of the Mediterranean climate and its applicability in many other countries in the region. Furthermore, Algerian public policy prioritises energy efficiency as a matter of urgency [20]. The weather profile from Sahabi Abed and Matzarakis [21] for the region of Algiers was used for the base case model, which was provided to the design experiment participants, in addition to input in DesignBuilder, for the simulations and, subsequently, for the optimisation.
3.2.2. Base Case Model Description
The distribution of energy usage in Algeria by sector indicates that buildings account for a very high proportion. The residential sector in particular accounts for 43% of the total energy consumption, according to the energy balance released by [22]. For this reason, a design task base case model of a residential building was chosen for this design exercise. The model was a typical five-storey residential block, with each floor comprised of four apartments, as shown in the floorplans of Figure 3 and Figure 4, the latter of which shows a 3D visualisation of the whole building.
The floor area of each floor is 511.55 m2 with a floor-to-ceiling height of 3.0 m. The base case building was assumed to be free from overshadowing from surrounding buildings.
3.2.3. Sampling Strategy
Purposive sampling was used to recruit architects for the design experiment. This approach is suited to understanding aspects of real-world architecture practice. As years of experience was identified as a significant variable, the sample was stratified by recruiting roughly uniform numbers of participants in each of the five-year experience ranges. In Algeria, as is the case for most countries, the architect typically leads the entire building design development process. In early design, architects receive relevant project requirement information from a developer or owner and develop a conceptual building design. Fundamental decisions are made at this stage, such as 3D massing, building orientation, and materials.
3.2.4. Questionnaire Survey to Identify Pertinent Design Variables
Design experiment participants would be asked to make design decisions about particular building characteristics. For this purpose, it was necessary to identify the most important design variables affecting building energy performance. Therefore, an online questionnaire survey was issued to a separate sample of practising architects to help set up the design experiment. Architects were asked to rank a number of pertinent variables identified from the literature by order of importance to achieving an energy-efficient building and then to rate the potential impact of each variable on final energy performance. A set of 109 questionnaire responses was received, and the data can be found in Figure 5. Based on these assessments of importance and impact by professional architects, the following variables were selected to be incorporated into the design exercise: external wall construction, roof construction, glazing type, window frame type, south/east/west shading, building orientation, and window-to-wall ratio.
Once the base case model was fully developed and pertinent design variables were identified, a design exercise was developed in which professional architects as research participants were asked to modify the base design by altering the pertinent design variables in order to balance the two objectives of construction cost and energy efficiency. Each design generated by a participant had to be assessed for its achievement of the two design objectives. DesignBuilder was selected as the design simulation application, as it has an interface for a leading global simulation engine, EnergyPlus 9.6.0.
Two hundred participants were issued a design exercise leaflet along with the base case model of the residential building block in RVT (Autodesk Revit 2019) format. The data were collected in dual format: in-person visits to architectural offices, where possible, and online, where in-person visits were not feasible.
Participants were instructed to improve the base case model’s performance against the two design objectives by reconsidering the pertinent design variables. The list of the design variables was created as a database within Autodesk Revit. The term ‘design variables’ in this study refers to those variables that the designer can change. This is consistent with other studies, such as Hamdy et al., [23], and Sasena et al. [24]. The term ‘design choices’ refers to the values that each design variable (in this case, window type) may take, e.g., single glazing, double, grazing, and triple glazing. Each design variable had a number of design choices associated with it, and the architects had to manually apply the design choices in the Revit model provided. These parameters are laid out in Table 1, Table 2 and Table 3. The types of external wall construction were sourced from the study of Baglivo et al. [25] on optimised external walls in the Mediterranean climate. Each type of wall has predefined layers in which different materials (in the layer thickness range column) are allowed to vary according to the design variables and choices that architects think are the most appropriate. The thermal specifications of insulation materials, in terms of thermal conductivity, heat capacity, and density, have been identified based on the standard EN ISO 10456 [26].
All the architects’ generated designs were simulated in DesignBuilder, taking into account all the configurations decided by the participants; this includes the type of wall, insulation thickness layers, type of roof, type of glazing and window frame, building orientation, window-to-wall ratio, air change rate and local shading devices, and the size. The purpose of this sub-task was to evaluate the two MOPs of each design, which were energy consumption in kWh/m2 and total construction cost in DZD, as shown in Figure 6.
Following the selection by each architect of alternative design options in line with the two design objectives, the resulting design was assessed by using DesignBuilder to predict its energy consumption and construction cost, as measures of performance (MOP). The gbXML standard was used to export the models from Revit and import them into DesignBuilder. A simulation was run to calculate the expected lifecycle energy consumption in kWh/m2 and total construction cost in DZD (Algerian dinar).
A few assumptions had to be made in DesignBuilder when entering the simulation input settings:
- Activity: The average occupancy for residential apartments in Algeria is four people per apartment. A typical family would consist of a father, mother, and two to three children. From the building layout, the total occupancy area of each flat is approximately 90 m2. Therefore, the occupancy density specified in the simulation was 0.055 people/m2. Furthermore, DesignBuilder’s default metabolic factors were used: 1.00 for men, 0.85 for women, and 0.75 for children or an average value if there is a mix of occupants.
- The lighting heat gain was set to 7 W/m2 from the CIBSE energy code of practice [27]. The default diversity factor of 70% was used for the lighting heat gain, since it is unlikely that all lights will be on simultaneously. From the assumed occupancy schedule, lighting in the base case model was set to be used daily from 12:00 to 23:00.
- The heating and cooling setpoint temperatures were selected as default values; 21 °C for heating and 25 °C for cooling.
- Minimum fresh air was assumed to be 10 L/s–person, as suggested in the CIBSE 2015 guide [27]
- Depending on the nature of the treated zone in the energy simulation model, the internal gains, such as from occupants and equipment, were left at their default values from DesignBuilder, depending on the nature of the occupant activity schedule and metabolic rate.
Weather data that are used in DesignBuilder are based on EnergyPlus weather files that define external conditions during simulation. Each location has a separate weather file describing the external temperature, atmospheric conditions, solar radiation, etc., for every hour of the year at that location. The file location of this research was set in the North African region of Algeria and specifically in the capital Algiers.
3.3. Computer Optimisation
DesignBuilder’s dynamic simulation and optimisation was used to compare the performance of the human-generated designs to those from computer-generated optimum design. Computer optimisation was conducted to benchmark the quality of the human designs, highlighting the role of human reasoning. Optimisation in design is a process where design variables and objectives need to be clearly defined in the early stages of the design process [28]. The variation and the number of design variables make it possible to identify a large number of possible designs including the interactions of the design variables that eventually determine how buildings are constructed and use energy [29]. The more design variables there are and the more choices available for each design variable, the larger the design search space (the set of all possible designs). DesignBuilder has a built-in genetic algorithm in its optimisation engine called r It is widely used as a multi-objective method that supports the search for a good trade-off between a well-converged and well-distributed solution set. This method is used in this research to search for optimal design solutions. The same entire set of design variables and design choices that were used for the human architect design experiment were also used in the optimisation exercise to allow a robust comparison between the human architects’ generated design and the computation. The design variables were taken through a wide range to ensure that the optimum design could be found in that range. The design variables were incremented in sufficiently small steps through their respective ranges to avoid missing optimum values. The variables used in this experiment fall into two categories: variables with numerical design choices and those with non-numerical design choices. The ranges and incremental step values for the numerical variables were set heuristically, informed by best practice in design. For non-numerical variables such as building floor and roof, only one design alternative option was set to limit the sheer number of design possibilities.
4. Results and Discussion
The results of the preliminary survey are shown in Figure 5. These design variables emerged from the literature and are assumed to be within the architect’s control. The aim was to rank the variables by order of importance, as a precursor to including the most important variables in the subsequent design experiment. However, it is recognised that architects might not have full freedom to make early-stage decisions due to high uncertainty and the complexity of stakeholders. Therefore, architects were also asked to rate the impact on energy performance of the variables [7]. External wall construction was ranked as the most important variable with an impact factor of four, followed by roof construction and window-to-wall ratio (WWR), each with an impact weight factor of three. The variables included in the design exercise (Table 1, Table 2 and Table 3) were carefully selected to represent the top ranked variables from Figure 5. The design exercise was delivered to a sample of 200 professional architects. A total of 134 valid design solutions out of 200 were received. The location of the study had a direct effect on the results due to prevalent common design practices and design culture.
The building was modelled in DesignBuilder, considering all the configurations decided by the participants based on Table 1, Table 2 and Table 3. The DesignBuilder platform has a construction cost component which was used to calculate the initial construction cost of each design variant provided by the architects. The cost model used in the calculation is based on per unit area averages from a published database linked to DesignBuilder. The prices of materials and labour cost were sourced from ©CYPE Ingenieros, S.A [30]. The price generator is an engine that allows a calculation of the real cost of construction materials and labour in Algeria, where the experiment was carried out. It provides an updated real-time construction cost adjusted to the market in which the project is built. A cost correction factor was applied specifically for Algeria. The accuracy of absolute cost estimates is unimportant here, but rather the relative costs used to compare design alternatives and identify optimum solutions were used instead. Table 4 shows the results of the two performance indicators, energy consumption in kWh/m2/year and construction cost in DZD for the 134 designs submitted by the architects.
The results for the designs generated by DesignBuilder’s optimisation engine are shown in the optimisation converged at 309 model iterations, of which 13 were identified as optimal and formed a Pareto front, as shown in red in Figure 7. The Pareto front is identified with DesignBuilder optimisation automatically, and it uses equal weights for each of the design objectives. In design, there is always a trade-off between competing design objectives. Some of these design objectives need to be given a certain priority weight, each of which makes those objectives that are given a higher priority in order to drive the design. Wang et al. [31] reported that, in practice, the use of equal weights (50/50) is the most popular weighting method.
Table 5 summarises the Pareto front solutions.
The optimisation analysis was carried out on the basis of cross-referencing all design variables from Table 1, Table 2 and Table 3. The Pareto solutions represent the designs that achieved the minimum values for the energy and cost objectives, and neither objective could be improved without harming the other. For both design objectives, minimisation is desired.
Among all the solutions generated by the optimisation, the Pareto curve comprises 13 solutions, as shown in Figure 7. For each Pareto optimum, the performance in the energy consumption objective could only be improved by decreasing the performance in construction cost.
When examining the Pareto front curve more closely, there are optimal solutions that can comparatively be more favourable than others. For instance, from Table 5, Pareto 08 achieved the most energy-efficient design with 89.86 kWh/m2/year. This Pareto optimal is also the most expensive design on the Pareto front. Given that Pareto 08 offers better outcomes at the same cost, we can confidently state that it dominates Pareto 05. In the language of Pareto efficiency, this means that Pareto 08 is strictly preferred over Pareto 05. Therefore, Pareto 08 is used as the baseline, against which the results from test subjects are compared.
Figure 8 shows a comparison of the architects’ designs versus the designs that were generated with multi-objective optimisation in DesignBuilder. The first observation is that the bulk majority of architects who took part in this study are far from the range of the optimal designs generated by the optimisation. The human-generated designs can be considered failed designs as they do not achieve a trade-off between the two competing design objectives. The primary energy consumption target was set to 100 kWh/m2/yr in the design exercise, yet almost none of the architects were able to come close this energy figure.
Computer optimisation clearly outperforms humans when searching for the best trade-off between energy performance and construction cost. The architects had generated various designs at substantially different performance levels when a considerably cheaper design at a much lower energy consumption was found using optimisation. The analysis of the human designs suggests that the architects are, to a certain extent, not concerned with energy efficiency. Architects seem to struggle to position energy efficiency within their professional boundaries. Figure 9 categorises architects’ performance by years of experience and suggests that the relationship between design performance and experience is not decisive. The relationship is a little clearer from Figure 10 and Figure 11. Figure 10 shows that the energy consumption of human-generated designs rises slightly as the architects’ years of experience rises. Figure 11 shows that more experienced architects generated cheaper designs. The error bars shown in the charts suggest that these differences might not be statistically significant.
Architects with between 0 and 5 years of experience generated the most energy efficient designs, while those having over 20 years of experience generated the least energy efficient designs. The failure in achieving a competitive trade-off between design objectives worsens as experience increases. The mean energy consumption and mean construction cost was calculated for the 13 solutions in the Pareto front and taken to represent the performance of the computer optimisation. Figure 11 indicates this mean energy consumption for the optimization Pareto set so that it can be compared to that for architects in various ranges of experience. The computer optimisation decisively outperforms even the best-performing group of architects, those with the least experience.
The findings of this study suggest that professional design experience only improves construction cost prediction but not design for low energy consumption. From Figure 9, it appears that only a subgroup of architects was able to compete with the optimisation in terms of construction cost (the horizontal strip of points at about the DZD 230 million level, equivalent to GBP 1,336,518). The construction cost performance of the architects’ designs improves (i.e., cost decreases) as the architects become more experienced and appears to converge towards the optimisation prediction, as seen in Figure 10. (The least two categories of experience on the left-hand side of Figure 10 are a minor exception, but these might not be significant.) On the other hand, the energy performance of the architects’ designs deteriorates (i.e., predicted energy consumption rises) as the architects become more experienced, as can be seen in Figure 11.
It is striking in the results that the human architects (despite being instructed explicitly to balance the two objectives of construction cost and energy efficiency), were more concerned with construction cost. The decline in energy performance as the architects become more experienced is also striking. Construction cost estimation is known to be highly affected by the architects’ reasoning that is nurtured through learning by doing and professional expertise [32]. Elfaki and Alatawi [33] argue that early construction cost estimation relies on human judgement, and there often are differences between estimates even in the same project. This resonates with the data from the design experiment. Despite receiving the same data, base case model, and instructions, there was a large variance in the cost performance of the human architects (as evidenced by the large error bars in Figure 10). It appears, however, that the more experienced architects can compete with the Pareto front in terms of construction cost estimation.
In his book on Building Performance Analysis, De Wilde [6], on page 457, contrasts the research into performance-based design with the practice of performance-based design decision-making. He observes that, whereas the research (often by academics) assumes a normative model of decision-making, design practice is actually based on tacit knowledge, where decisions are made “on the go”. Performance criteria are usually incorporated into designs through the application of tacit knowledge and expertise. This aligns with the research of Woo et al. [34], who similarly highlights the importance of tacit (over explicit) knowledge in the design and construction of buildings. From the research reported here, the central issue seems to be the balance accorded to the two competing objectives of construction cost and energy performance. The knowledge acquired during education is effective in achieving better energy-efficient designs, but experiential knowledge seems to tip the balance more in favour of reducing construction cost.
5. Conclusions
In general, the contribution of this research highlights that computer optimisation outperforms human architects decisively in terms of the energy performance of the designs but less conclusively in terms of the cost performance of the designs. Rather than improve design performance in general, the architects’ reasoning that is nurtured through experience seems to tip the balance from energy performance to cost performance. That is to say, while computer optimisation shows promise in cost-effective design, the results are less conclusive when compared to human architects. This nuanced outcome can be explained by that the fact that experienced architects bring invaluable insights into local building practices, material availability, and labour costs. Furthermore, in terms of qualitative factors, humans excel at considering subjective elements like aesthetics and user experience, which can impact overall project value.
The findings in this study challenge the assumption that experienced architects possess extensive and effective reasoning in making design decisions. The reliance on the designers’ reasoning in this context is only useful when making construction cost predictions. Considering the objective of energy-efficient design in isolation, the performance of novice architects is good (perhaps an indication of the Algeria system of architecture training) but still sub-optimal compared to the computer optimisation. Practical experience causes it to deteriorate further. In Algeria, during the first two years after graduation, architects undertake training within a qualified architecture practice before being accredited to practise independently. During this period, there is a knowledge transfer from expert architects to novices, as novices collaborate on projects with experienced architects. This allows experienced architects to share their knowledge with novices through “Communities of Practice”. This shared knowledge is fundamental to the practice of architecture [34]. Moreover, this can also explain that, particularly among the younger population, there is a growing ecological awareness driving a shift towards sustainability. This demographic is increasingly passionate about addressing climate change through more environmentally responsible design decisions. However, further research is needed to understand why mid-career professional practice of architecture seems to erode the importance of energy efficiency which is imparted during architecture training.
There is scope for future research in understanding the inner workings of the human designers’ reasoning during design decision-making, perhaps using think-aloud verbal protocols. The research reported here was limited in its use of a single design problem, a single geographic location, and a single computer optimisation platform. A larger sample size would have provided more statistical power. Ultimately, the framing of human versus computer might be unhelpful. Architects can deploy computer optimisation in a meaningful way appropriate to the respective strengths of humans and computers. Computer optimisation can augment human reasoning, rather than compete with it or replace it. This is particularly true in creative design or design teamwork [35].
Author Contributions
Methodology, S.A.O. and M.E.; Writing—review & editing, M.E.; Supervision, P.D. All authors have read and agreed to the published version of the manuscript.
Funding
This research received no external funding.
Institutional Review Board Statement
The study was conducted in accordance with the Declaration of Helsinki and approved by the Institutional Review Ethics Committee of Loughborough University (protocol code DT_5329 on 8 December 2017).
Informed Consent Statement
Informed consent was obtained from all subjects involved in the study.
Data Availability Statement
Data are contained within the article.
Conflicts of Interest
The authors declare no conflict of interest.
References
- Mahmoud, R.; Kamara, J.M.; Burford, N. Opportunities and Limitations of Building Energy Performance Simulation Tools in the Early Stages of Building Design in the UK. Sustainability 2020, 12, 9702. [Google Scholar] [CrossRef]
- Terim Cavka, B.; Cavka, H.B.; Salehi, M.M. An Investigation of the Design Process’s Effect on a High-Performance Building’s Actual Energy System Performance. J. Integr. Des. Process Sci. 2022, 26, 85–100. [Google Scholar] [CrossRef]
- Attia, S.; Gratia, E.; De Herde, A.; Hensen, J.L.M. Simulation-Based Decision Support Tool for Early Stages of Zero-Energy Building Design. Energy Build. 2012, 49, 2–15. [Google Scholar] [CrossRef]
- Tresidder, E. Accelerated Optimisation Methods for Low-Carbon Building Design. Ph.D. Thesis, De Montfort University, Leicester, UK, 2014. [Google Scholar]
- Ercan, B.; Elias-Ozkan, S.T. Performance-Based Parametric Design Explorations: A Method for Generating Appropriate Building Components. Des. Stud. 2015, 38, 33–53. [Google Scholar] [CrossRef]
- de Wilde, P. Building Performance Analysis, 1st ed.; Wiley-Blackwell: Hoboken, NJ, USA, 2018. [Google Scholar]
- Urban, B.J. The MIT Design Advisor: Simple and Rapid Energy Simulation of Early-Stage Building Designs. Ph.D. Thesis, Massachusetts Institute of Technology, Cambridge, MA, USA, 2007. [Google Scholar]
- Rezaee, R.; Brown, J.; Haymaker, J.; Augenbroe, G. A Novel Inverse Data Driven Modelling Approach to Performance-Based Building Design during Early Stages. Adv. Eng. Inform. 2019, 41, 100925. [Google Scholar] [CrossRef]
- Hirschi, N.W.; Frey, D.D. Cognition and Complexity: An Experiment on the Effect of Coupling in Parameter Design. Res. Eng. Des. 2002, 13, 123–131. [Google Scholar] [CrossRef]
- Daly, S.R.; Yilmaz, S.; Christian, J.L.; Seifert, C.M.; Gonzalez, R. Design Heuristics in Engineering Concept Generation. J. Eng. Educ. 2012, 101, 601–629. [Google Scholar] [CrossRef]
- Dinar, M.; Shah, J.J.; Cagan, J.; Leifer, L.; Linsey, J.; Smith, S.M.; Hernandez, N.V. Empirical Studies of Designer Thinking: Past, Present, and Future. J. Mech. Des. 2015, 137, 021101. [Google Scholar] [CrossRef]
- Dorst, K. The Core of “design Thinking” and Its Application. Des. Stud. 2011, 32, 521–532. [Google Scholar] [CrossRef]
- Connell, N.A.D.; Klein, J.H.; Powell, P.L. It’s Tacit Knowledge but Not as We Know It: Redirecting the Search for Knowledge. J. Oper. Res. Soc. 2003, 54, 140–152. [Google Scholar] [CrossRef]
- Petersen, S. General Rights Simulation-Based Support for Integrated Design of New Low-Energy Office Buildings. Ph.D. Thesis, Technical University of Denmark, Department of Civil Engineering, Kongens Lyngby, Denmark, 2011. [Google Scholar]
- Stanitsa, A.; Hallett, S.H.; Jude, S. Investigating Key Factors Influencing Decision-Making in the Design of Buildings and Places: A Survey of Stakeholders’ Perception. Archit. Struct. Constr. 2022, 2, 381–401. [Google Scholar] [CrossRef]
- Singh, H.; Cascini, G.; McComb, C. Comparing Design Outcomes Achieved by Teams of Expert and Novice Designers through Agent-Based Simulation. Proc. Des. Soc. 2021, 1, 661–670. [Google Scholar] [CrossRef]
- Björklund, T.A. Initial Mental Representations of Design Problems: Differences between Experts and Novices. Des. Stud. 2013, 34, 135–160. [Google Scholar] [CrossRef]
- Moradi, A.; Kavgic, M.; Costanzo, V.; Evola, G. Impact of Typical and Actual Weather Years on the Energy Simulation of Buildings with Different Construction Features and under Different Climates. Energy 2023, 270, 126875. [Google Scholar] [CrossRef]
- Kalvelage, K.; Passe, U.; Rabideau, S.; Takle, E.S. Changing Climate: The Effects on Energy Demand and Human Comfort. Energy Build. 2014, 76, 373–380. [Google Scholar] [CrossRef]
- Seddiki, M.; Bennadji, A.; Tehami, M. Barriers to the Adoption of Energy Efficiency Measures in Mostaganem, Algeria. J. Constr. Dev. Ctries. 2020, 25, 39–61. [Google Scholar] [CrossRef]
- Sahabi Abed, S.; Matzarakis, A. Seasonal Regional Differentiation of Human Thermal Comfort Conditions in Algeria. Adv. Meteorol. 2017, 2017, 9193871. [Google Scholar] [CrossRef]
- Errebai, F.B.; Derradji, L.; Limam, A.; Amara, M. Energy Efficiency and Economic Profitability between Standard and High Energy Performance Dwellings in a Semi-Arid Climate. Int. J. Ambient. Energy 2022, 43, 4020–4036. [Google Scholar] [CrossRef]
- Hamdy, M.; Palonen, M.; Hasan, A. Implementation of Pareto-Archive Nsga-Ii Algorithms to a Nearly-Zero-Energy Building Optimisation Problem. In Proceedings of the BSO Conference 2012: 1st Conference of IBPSA-England, Loughborough, UK, 10–11 September 2012; IBPSA: Loughborough, UK, 2012; pp. 181–188. [Google Scholar]
- Sasena, M.J.; Papalambros, P.; Goovaerts, P. Exploration of Metamodeling Sampling Criteria for Constrained Global Optimization. Eng. Optim. 2002, 34, 263–278. [Google Scholar] [CrossRef]
- Baglivo, C.; Congedo, P.M.; Fazio, A.; Laforgia, D. Multi-Objective Optimization Analysis for High Efficiency External Walls of Zero Energy Buildings (ZEB) in the Mediterranean Climate. Energy Build. 2014, 84, 483–492. [Google Scholar] [CrossRef]
- ISO 10456:2007; Building Materials and Products—Hygrothermal Properties—Tabulated Design Values and Procedures for Determining Declared and Design Thermal Values. International Organization for Standardization: Geneva, Switzerland, 2007.
- Parand, F.; Nicol, F.; Nikolopoulou, M. Environmental Design. (Incorporating Corrections as of May 2019); Chartered Institution of Building Services Engineers: London, UK, 2019. [Google Scholar]
- Negendahl, K.; Nielsen, T.R. Building Energy Optimization in the Early Design Stages: A Simplified Method. Energy Build. 2015, 105, 88–99. [Google Scholar] [CrossRef]
- Hsu, D. Identifying Key Variables and Interactions in Statistical Models Ofbuilding Energy Consumption Using Regularization. Energy 2015, 83, 144–155. [Google Scholar] [CrossRef]
- CYPE Ingenieros; CYPE Ingenieros, S.A.: Alicante, Spain, 2024.
- Wang, B.; Zhou, J.; Koschny, T.; Kafesaki, M.; Soukoulis, C.M. Chiral Metamaterials: Simulations and Experiments. J. Opt. A Pure Appl. Opt. 2009, 11, 114003. [Google Scholar] [CrossRef]
- Du, R.K.J.; El-Gafy, M. Human Computation Enabled Organizational Learning in the Face of Deep Uncertainty: Example of Conceptual Estimating. J. Inf. Technol. Constr. (ITcon) 2015, 20, 275–294. [Google Scholar]
- Elfaki, A.O.; Alatawi, S. Representing the Knowledge of Public Construction Project Cost Estimator by Using Rule-Based Method. J. Build. Constr. Plan. Res. 2015, 03, 189–195. [Google Scholar] [CrossRef]
- Woo, J.H.; Clayton, M.J.; Johnson, R.E.; Flores, B.E.; Ellis, C. Dynamic Knowledge Map: Reusing Experts’ Tacit Knowledge in the AEC Industry. Autom. Constr. 2004, 13, 203–207. [Google Scholar] [CrossRef]
- Zhang, G.; Raina, A.; Cagan, J.; McComb, C. A Cautionary Tale about the Impact of AI on Human Design Teams. Des. Stud. 2021, 72, 100990. [Google Scholar] [CrossRef]
Figure 1.
Research method.
Figure 2.
Experimental procedure for design experiments.
Figure 3.
Typical architectural floorplan of the base case model.
Figure 4.
Entire building 3D visualization of the base case model considered for this study.
Figure 5.
Design variable ranking by order of importance in bars and impact average weight by line design exercise.
Figure 5.
Design variable ranking by order of importance in bars and impact average weight by line design exercise.
Figure 6.
Workflow of model’s preparation for energy and cost performance simulation in DesignBuilder.
Figure 6.
Workflow of model’s preparation for energy and cost performance simulation in DesignBuilder.
Figure 7.
Trade-off between energy consumption and cost using computer optimisation (The red represent the front that is referred to as Pareto Front to distinguish the best trade off design options from the rest of the design population).
Figure 7.
Trade-off between energy consumption and cost using computer optimisation (The red represent the front that is referred to as Pareto Front to distinguish the best trade off design options from the rest of the design population).
Figure 8.
Computer-generated Pareto optimal designs versus human designs, plotted by two competing design objectives of construction cost and energy consumption.
Figure 8.
Computer-generated Pareto optimal designs versus human designs, plotted by two competing design objectives of construction cost and energy consumption.
Figure 9.
Trade-off between energy consumption and construction cost according for architects with various years of experience of the architects.
Figure 9.
Trade-off between energy consumption and construction cost according for architects with various years of experience of the architects.
Figure 10.
Mean construction cost of designs produced by architects with various years of experience.
Figure 10.
Mean construction cost of designs produced by architects with various years of experience.
Figure 11.
Mean energy consumption of design produced by architects with various years of experience.
Figure 11.
Mean energy consumption of design produced by architects with various years of experience.
Table 1.
Building fabric variables.
| Choices for Building Envelope Variable | Layers | Choices for Layer Thickness Choices (mm) |
|---|---|---|
| Types of external walls; the choice of internal/external insulation was left to the participants | ||
| External heavyweight wall Type 1 | Concrete | 90/100/110/120/130/140/150 |
| Polyurethane foam 1 | 20/30/40/50/60/70/80/90/100/120 | |
| Concrete exp. clay | 50 | |
| Polyurethane foam 2 | 20/30/40/50/60/70/80/90/100/110/120/130/140 | |
| Concrete | 100 | |
| External heavyweight wall Type 2 | Concrete | 10/20/30/40/50/60/70/80/90/100/110/120/130/140/150 |
| Cork panel exp. | 10/20/30/40/50/60/80/100/120/140/160 | |
| Cellulose fibre | 30/40/50/60/80/100/120/140/160/180 | |
| Brick | 120 | |
| Polyethylene exp. | 2/3/4/5/6/8/10/12/15/20/25/30 | |
| External heavyweight wall Type 3 | Concrete | 90/100/110/120/130/140/150 |
| Polyurethane foam 1 | 20/30/40/50/60/70/80/90/100/120 | |
| Cross-laminated timber panels | 60/78/95/128/146/162/202 | |
| Polyurethane foam 2 | 20/30/40/50/60/70/80/90/100/110/120/130/140 | |
| Plaster | 15 | |
| External lightweight wall Type 4 | Concrete | 90/100/110/120/130/140/150 |
| Hemp fibre | 30/40/50/60/80/100/120/140/160/180/200/220/240 | |
| Air | 5/10/15/20/25/30/35/40/45/50 | |
| Cross-laminated timber panels | 60/78/95/128/146/162/202 | |
| Cork panel exp. | 10/20/30/40/50/60/80/100/120/140/160 | |
| External lightweight wall Type 5 | Concrete | 90/100/110/120/130/140/150 |
| Fibreboard | 20/40/60/80/100/120/140/160 | |
| Polyethene exp. | 2/3/4/5/8/10/12/15/20/25/30 | |
| Cork panel exp. | 10/20/30/40/50/60/80/100/120/140/160 | |
| Plaster | 15 | |
| External lightweight wall Type 6 | Concrete | 90/100/110/120/130/140/150 |
| Wood fibre hardboard | 20/40/60/80/100/120/140/160 | |
| Wood fibre hardboard | 20/40/60/80/100/120/140/160 | |
| Polyurethane foam 1 | 20/30/40/50/60/70/80/90/100/120 | |
| OSB (Oriented Strand Board) | 15/18/22/25 | |
| Roof type | ||
| Ground floor This part is fixed | Ceramic finishing | 7 |
| Cement mortar | 50 | |
| Polystyrene | 0 to 50 | |
| Reinforced concrete slab | 200 | |
| Roof type 1-hollow core | Plasterboard | 15 |
| Hollow concrete slab | 160 | |
| Compression slab | 40 | |
| Polystyrene | 0 to 180 | |
| Sealing layer | 20 | |
| Roof type 2-prestressed reinforced concrete | Plasterboard | 15 |
| Solid slab | 150 | |
| Vapour barrier | 20 | |
| Polystyrene | 0 to 180 | |
| Sealing layer | 20 | |
Table 2.
Characteristics of glazing typologies.
| Window Variables | Window Layer Thickness (mm) | U-Value (W/m2 K) | SHGC | Solar Transmission | Solar Reflectance | Visual Transmission |
|---|---|---|---|---|---|---|
| Single glazing (Glz1) | 2.5 | 5.74 | 0.87 | 0.85 | 0.075 | 0.901 |
| Double glazing filled with air (Glz2) | 2.5/12.7/2.5 | 2.95 | 0.777 | 0.727 | 0.129 | 0.817 |
| Triple glazing filled with air (Glz3) | 2.5/12.7/2.5/12.7/2.5 | 2 | 0.7 | 0.624 | 0.168 | 0.744 |
| Double glazing filled with air +1 low emissivity layers facing inward (Glz4) | 3/12.7/2.5 | 1.76 | 0.597 | 0.544 | 0.22 | 0.769 |
| Double glazing filled with argon +1 low emissivity layers facing inward (Glz5) | 3/12.7/2.5 | 1.43 | 0.596 | 0.544 | 0.22 | 0.769 |
Table 3.
Different levels of selected parameters.
| Other Variables | Symbol | Lower Limit | Upper Limit | Unit |
|---|---|---|---|---|
| Shading coefficient of windows facing south | SCs | 0 | 1 | - |
| Shading coefficient of windows facing east | SCe | 0 | 1 | - |
| Shading coefficient of windows facing west | SCw | 0 | 1 | - |
| Glazing | Glz | Glz1, Glz2, Glz3, Glz4, Glz5 | - | - |
| Type of window frame | WF | Aluminium, Wood, UPVC | - | - |
| Building Orientation | BO | ±45°, ±90° | - | Degrees from North |
| Window-to-wall ratio | WWR | 20 | 90 | % |
Table 4.
Simulation results of energy consumption and construction cost for architects’ designs.
| Human Designs | Energy kWh/m2/Year | Construction Cost in Euros |
|---|---|---|
| Architect 1 | 143.07 | 260,210,000.00 |
| Architect 2 | 151.67 | 257,351,000.00 |
| Architect 3 | 150.01 | 253,508,000.00 |
| Architect 4 | 173.47 | 253,367,000.00 |
| Architect 5 | 172.64 | 262,919,000.00 |
| Architect 6 | 157.38 | 257,025,000.00 |
| Architect 7 | 182.19 | 254,735,000.00 |
| Architect 8 | 174.04 | 260,942,000.00 |
| Architect 9 | 156.81 | 261,224,000.00 |
| Architect 10 | 214.39 | 232,977,000.00 |
| Architect 11 | 215.87 | 232,580,000.00 |
| Architect 12 | 190.56 | 235,461,000.00 |
| Architect 13 | 163.46 | 260,451,000.00 |
| Architect 14 | 190.37 | 235,731,000.00 |
| Architect 15 | 215.61 | 232,501,000.00 |
| Architect 16 | 136.03 | 260,651,000.00 |
| Architect 17 | 137.92 | 260,965,000.00 |
| Architect 18 | 215.85 | 232,117,000.00 |
| Architect 19 | 213.73 | 236,004,000.00 |
| Architect 20 | 170.91 | 260,738,000.00 |
| Architect 21 | 213.09 | 235,553,000.00 |
| Architect 22 | 169.02 | 262,370,000.00 |
| Architect 23 | 169.5 | 261,038,000.00 |
| Architect 24 | 169.98 | 260,500,000.00 |
| Architect 25 | 215.3 | 232,201,000.00 |
| Architect 26 | 156.01 | 261,224,000.00 |
| Architect 27 | 173.29 | 260,320,000.00 |
| Architect 28 | 146.72 | 257,943,000.00 |
| Architect 29 | 215.37 | 232,820,000.00 |
| Architect 30 | 143.01 | 260,222,000.00 |
| Architect 31 | 143.02 | 260,222,000.00 |
| Architect 32 | 151.1 | 252,390,000.00 |
| Architect 33 | 173.1 | 253,239,000.00 |
| Architect 34 | 170.7 | 262,200,000.00 |
| Architect 35 | 161.04 | 257,285,000.00 |
| Architect 36 | 171.3 | 262,416,000.00 |
| Architect 37 | 155.23 | 257,001,000.00 |
| Architect 38 | 214.85 | 232,773,000.00 |
| Architect 39 | 143.1 | 230,201,000.00 |
| Architect 40 | 153.69 | 254,490,000.00 |
| Architect 41 | 150.29 | 252,309,000.00 |
| Architect 42 | 172.96 | 254,765,000.00 |
| Architect 43 | 173.79 | 262,919,000.00 |
| Architect 44 | 156.1 | 289,391,000.00 |
| Architect 45 | 180.2 | 250,731,000.00 |
| Architect 46 | 169.04 | 259,840,000.00 |
| Architect 47 | 154.82 | 250,443,000.00 |
| Architect 48 | 214.2 | 232,983,000.00 |
| Architect 49 | 215.9 | 232,490,000.00 |
| Architect 50 | 196.8 | 239,309,000.00 |
| Architect 51 | 160.01 | 259,099,000.00 |
| Architect 52 | 193.4 | 234,870,000.00 |
| Architect 53 | 213.21 | 232,601,000.00 |
| Architect 54 | 132.1 | 261,390,000.00 |
| Architect 55 | 137.92 | 260,965,000.00 |
| Architect 56 | 215.7 | 232,877,000.00 |
| Architect 57 | 213.73 | 232,004,000.00 |
| Architect 58 | 170.91 | 260,738,000.00 |
| Architect 59 | 213.09 | 232,553,000.00 |
| Architect 60 | 169.02 | 262,370,000.00 |
| Architect 61 | 167.6 | 261,038,000.00 |
| Architect 62 | 140.1 | 260,210,000.00 |
| Architect 63 | 149.9 | 256,781,000.00 |
| Architect 64 | 151.57 | 256,951,000.00 |
| Architect 65 | 170.91 | 255,547,000.00 |
| Architect 66 | 171.15 | 262,919,000.00 |
| Architect 67 | 159.4 | 253,580,000.00 |
| Architect 68 | 180.92 | 252,635,000.00 |
| Architect 69 | 173.89 | 259,998,000.00 |
| Architect 70 | 156.5 | 260,359,000.00 |
| Architect 71 | 215.28 | 232,500,000.00 |
| Architect 72 | 214.9 | 232,821,000.00 |
| Architect 73 | 191.12 | 235,461,000.00 |
| Architect 74 | 164.45 | 261,451,000.00 |
| Architect 75 | 190.37 | 235,319,000.00 |
| Architect 76 | 213.05 | 232,835,000.00 |
| Architect 77 | 130.94 | 261,903,000.00 |
| Architect 78 | 136.64 | 259,376,000.00 |
| Architect 79 | 213.58 | 232,769,000.00 |
| Architect 80 | 170.87 | 261,378,000.00 |
| Architect 81 | 142.64 | 256,644,000.00 |
| Architect 82 | 213.75 | 233,853,000.00 |
| Architect 83 | 141.52 | 261,913,000.00 |
| Architect 84 | 141.06 | 261,733,000.00 |
| Architect 85 | 214.83 | 232,809,000.00 |
| Architect 86 | 121.76 | 240,907,000.00 |
| Architect 87 | 159.3 | 264,600,000.00 |
| Architect 88 | 155.09 | 234,102,000.00 |
| Architect 89 | 169.51 | 260,333,000.00 |
| Architect 90 | 155.4 | 261,203,000.00 |
| Architect 91 | 159.41 | 258,901,000.00 |
| Architect 92 | 179.29 | 232,433,000.00 |
| Architect 93 | 170.09 | 261,222,000.00 |
| Architect 94 | 157.61 | 260,453,000.00 |
| Architect 95 | 215.31 | 232,430,000.00 |
| Architect 96 | 214.97 | 232,321,000.00 |
| Architect 97 | 187.69 | 234,200,000.00 |
| Architect 98 | 158.53 | 258,285,000.00 |
| Architect 99 | 187.43 | 235,416,000.00 |
| Architect 100 | 213.7 | 232,001,000.00 |
| Architect 101 | 134.79 | 234,773,000.00 |
| Architect 102 | 137.01 | 231,201,000.00 |
| Architect 103 | 213.92 | 232,490,000.00 |
| Architect 104 | 213.64 | 232,309,000.00 |
| Architect 105 | 171.03 | 252,765,000.00 |
| Architect 106 | 213.11 | 232,919,000.00 |
| Architect 107 | 170.03 | 285,391,000.00 |
| Architect 108 | 171.11 | 251,731,000.00 |
| Architect 109 | 139.94 | 260,840,000.00 |
| Architect 110 | 166.83 | 255,443,000.00 |
| Architect 111 | 214.48 | 231,983,000.00 |
| Architect 112 | 151.95 | 230,490,000.00 |
| Architect 113 | 176.19 | 240,309,000.00 |
| Architect 114 | 141.08 | 257,099,000.00 |
| Architect 115 | 213.74 | 235,870,000.00 |
| Architect 116 | 145.96 | 235,601,000.00 |
| Architect 117 | 141.95 | 262,390,000.00 |
| Architect 118 | 159.13 | 259,965,000.00 |
| Architect 119 | 179.19 | 234,877,000.00 |
| Architect 120 | 168.58 | 234,004,000.00 |
| Architect 121 | 156.19 | 258,738,000.00 |
| Architect 122 | 173.95 | 236,553,000.00 |
| Architect 123 | 144.02 | 262,370,000.00 |
| Architect 124 | 213.57 | 234,038,000.00 |
| Architect 125 | 149.21 | 258,210,000.00 |
| Architect 126 | 157.73 | 257,781,000.00 |
| Architect 127 | 149.28 | 253,951,000.00 |
| Architect 128 | 173.94 | 254,547,000.00 |
| Architect 129 | 167.02 | 261,919,000.00 |
| Architect 130 | 150.14 | 254,580,000.00 |
| Architect 131 | 179.21 | 251,635,000.00 |
| Architect 132 | 170.03 | 260,998,000.00 |
| Architect 133 | 151.82 | 261,359,000.00 |
| Architect 134 | 213.41 | 235,500,000.00 |
Table 5.
Energy consumption and construction cost for the optimal designs from the Pareto front.
| Pareto Design Iteration | Energy (kWh/m2/Year) | Cost (DZD) |
|---|---|---|
| Pareto 01 | 107.79 | DZD 212,521,000.00 |
| Pareto 02 | 95.43 | DZD 223,319,000.00 |
| Pareto 03 | 104.41 | DZD 212,935,000.00 |
| Pareto 04 | 103.4 | DZD 215,679,000.00 |
| Pareto 05 | 90.35 | DZD 230,570,000.00 |
| Pareto 06 | 101.37 | DZD 219,733,000.00 |
| Pareto 07 | 93.34 | DZD 226,477,000.00 |
| Pareto 08 | 89.86 | DZD 230,570,000.00 |
| Pareto 09 | 110.84 | DZD 207,874,000.00 |
| Pareto 10 | 95.77 | DZD 222,855,000.00 |
| Pareto 11 | 93.66 | DZD 226,012,000.00 |
| Pareto 12 | 101.95 | DZD 216,092,000.00 |
| Pareto 13 | 109.09 | DZD 211,031,000.00 |
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. |
© 2024 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/).
Share and Cite
MDPI and ACS Style
Ouldja, S.A.; Demian, P.; Eftekhari, M. The Impact of Architects’ Reasoning on Early Design Decision-Making for Energy-Efficient Buildings. Sustainability 2024, 16, 8220. https://doi.org/10.3390/su16188220
AMA Style
Ouldja SA, Demian P, Eftekhari M. The Impact of Architects’ Reasoning on Early Design Decision-Making for Energy-Efficient Buildings. Sustainability. 2024; 16(18):8220. https://doi.org/10.3390/su16188220
Chicago/Turabian StyleOuldja, Sid Ahmed, Peter Demian, and Mahroo Eftekhari. 2024. "The Impact of Architects’ Reasoning on Early Design Decision-Making for Energy-Efficient Buildings" Sustainability 16, no. 18: 8220. https://doi.org/10.3390/su16188220
APA StyleOuldja, S. A., Demian, P., & Eftekhari, M. (2024). The Impact of Architects’ Reasoning on Early Design Decision-Making for Energy-Efficient Buildings. Sustainability, 16(18), 8220. https://doi.org/10.3390/su16188220
Note that from the first issue of 2016, this journal uses article numbers instead of page numbers. See further details here.
