A Systematic Review of Multi-Objective Optimisation Building Energy Retrofit, with a Focus on Hot-Humid Climate Regions

Abstract

Globally, buildings are responsible for around 32% of energy consumption and 34% of greenhouse gas emissions. One reason for this is the poor energy efficiency of much of the current building stock. Around 75% of today’s buildings are projected to still be in use in 2050, highlighting the importance of retrofitting existing buildings for energy efficiency. Such a strategy presents substantial opportunities to decrease global energy consumption and greenhouse gas emissions. While building retrofit projects have been implemented in many developed countries, studies in hot-humid climates and developing countries are still lacking. The challenges posed by hot-humid climates make developing the right energy retrofit strategies even more difficult. This study reviews and analyses previous energy retrofit studies and optimisations in building energy retrofit that used multi-objective optimisation methods, especially in hot-humid climate regions, using a bibliometric mapping tool called “VOSviewer” (version 1.6.20). The study also follows the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) framework for systematic reviews. This literature review highlights the paucity of research related to Multi-Objective Optimisation building-energy retrofit for buildings in countries with hot-humid climates and aims to identify the optimal strategies for energy retrofitting buildings in hot-humid climates using an optimisation method. The results of this study will significantly impact stakeholders’ decision-making processes, enabling them to identify the most advantageous objectives and energy efficiency measures for retrofitting buildings.

1. Introduction

In 2023, the building sector accounted for approximately 32% of global energy demand and 34% of carbon dioxide (CO₂) emissions [1]. In hot-humid developing regions, such as Indonesia, buildings are estimated to consume around 30% of total national energy use, a proportion projected to increase to 40% by 2030 [2]. Many of the existing buildings have poor energy performance, which contributes to the large energy demand. Therefore, retrofitting the building stock represents a pivotal strategy to enhance energy efficiency, secure thermal comfort, improve occupant well-being and productivity, and reduce operational and maintenance expenditures. Moreover, building retrofits contribute to national energy security, reinforce corporate social responsibility commitments, and mitigate exposure to volatile energy prices. They also present broader socio-economic benefits, including job creation and the enhancement of urban liveability and functionality [3].

Building retrofitting refers to the systematic renovation of existing structures with the primary aims of enhancing energy efficiency, sustainability, and resilience. It comprises a broad spectrum of strategic interventions intended to improve building performance, reduce energy demand, and address structural deficiencies. The scholarly literature highlights the multidimensional nature of retrofitting, offering evidence-based guidance for its effective implementation and identifying critical factors influencing project success. Central to the retrofit process are energy audits and performance assessments, which enable the identification of opportunities and the prioritisation of energy-saving measures [4]. The overarching objectives of retrofit initiatives typically include optimising energy and environmental performance, reducing water consumption, improving indoor thermal comfort, and mitigating noise pollution, frequently achieved through the deployment of advanced technologies [5].

A range of energy retrofit measures is integral to enhancing the performance of existing buildings. First, improvements to the building envelope, through insulation, air sealing, and high-performance glazing, are essential for reducing energy consumption and improving thermal comfort [6,7]. Second, the optimisation of heating, ventilation, and cooling (HVAC) systems, including the adoption of high-efficiency models and advanced energy-saving technologies, can substantially lower energy demand in office buildings [6,7]. Third, lighting upgrades, particularly the installation of LED fixtures and sensor-based controls, contribute to significant energy savings [7]. Fourth, the integration of renewable energy technologies, such as photovoltaic (PV) systems, provides opportunities to offset energy consumption and advance sustainability objectives [7]. Fifth, advanced energy management systems enable optimised energy use, real-time performance monitoring, and the identification of further efficiency improvements [8,9]. Sixth, building intelligence enhancements, including the deployment of smart technologies and automation, facilitate more efficient operations and greater energy savings [10]. Seventh, structural retrofitting, which addresses deterioration and outdated building services, can improve both energy efficiency and indoor environmental quality [11]. Finally, comprehensive energy auditing and performance assessment play a critical role in diagnosing inefficiencies, informing retrofit strategies, and prioritising measures for implementation [6,12,13].

Despite substantial progress in retrofit research and implementation across developed regions, applications in hot-humid climates remain limited. Regulatory and market priorities in many such contexts still favour new construction, while existing buildings often lack systematic upgrade pathways. Yet, hot-humid environments present distinct optimisation challenges: high latent cooling loads, significant solar heat gains, reduced night-time cooling potential, and climate-driven requirements for shading, glazing performance, humidity control, and efficient HVAC operation. These conditions demand retrofit strategies that differ markedly from those in temperate or cold climates, where heating reduction and envelope insulation typically dominate.

Selecting appropriate retrofit measures is further complicated by the need to balance multiple, often competing objectives. Multi-objective optimisation (MOO) provides a structured approach to navigating these trade-offs by enabling simultaneous evaluation of energy use, carbon emissions, life-cycle cost, comfort, and other performance dimensions. Genetic algorithms, particularly the Non-dominated Sorting Genetic Algorithm II (NSGA-II), have become widely used tools in building retrofit optimisation, owing to their ability to explore complex search spaces and generate diverse Pareto-optimal solutions. Previous reviews have examined energy retrofit optimisation more broadly [14], yet they offer limited climate-specific insight and do not explicitly address the distinct priorities and performance drivers of hot-humid regions.

This review responds to that gap by offering the first systematic and comparative synthesis of multi-objective optimisation studies in building retrofit with specific attention to hot-humid climates. It combines a PRISMA-based systematic review with a bibliometric analysis to examine methodological trends, optimisation objectives, decision variables, and retrofit measures across different climatic contexts. A central aim is to clarify how retrofit optimisation in hot-humid regions differs from that in other climates in terms of dominant objectives (e.g., cooling energy reduction), influential design variables (e.g., glazing SHGC, shading geometry, cooling setpoints), and typical performance outcomes.

The contributions of this review are threefold. First, it identifies and analyses the small but growing body of MOO retrofit studies conducted in hot-humid climates, highlighting consistent strategies, performance patterns, and context-specific challenges. Second, it provides a cross-climate comparison to distinguish the optimisation behaviour of buildings in tropical climates from those in temperate, continental, and Mediterranean regions. Third, it critically evaluates the methodological implementation of MOO techniques, including the role of algorithm selection, simulation coupling, and objective formulation, to highlight strengths, limitations, and opportunities for future research.

Through this synthesis, the review aims to support researchers and practitioners in developing retrofit strategies that are responsive to the distinctive demands of hot-humid climates and in advancing the application of optimisation tools within these underexplored contexts.

2. Methodology

This review employs two complementary analytical components: a systematic review based on the PRISMA framework and a bibliometric analysis using VOSviewer. The systematic procedure for study selection, illustrated in Figure 1, follows the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) framework [15] and is structured into four stages: Identification, Screening, Eligibility, and Inclusion.

In the Identification stage, potentially relevant documents were collected using the search keywords listed in

Table 1. Search strategy keywords.

Topic	Keywords	Number of Documents	Documents in English (2010–2025)
Optimisation building retrofit	(optimization OR optimisation) AND retrofit AND building	1007	997
	(optimization OR optimisation) AND energy AND retrofit AND building	803	796
	(optimization OR optimisation) AND “energy retrofit” AND building	368	353

, drawing from Scopus, Web of Science, and Google databases. Before screening began, all retrieved records from Scopus, Web of Science, and Google Scholar were exported in the RIS format and consolidated into a single dataset. Database-generated metadata were harmonised to ensure consistent fields for title, authors, DOI, and publication year. Duplicate entries, primarily arising from overlap between Scopus and Web of Science, were identified and removed through a two-step process: (i) automated matching using reference management software, followed by (ii) a manual check to resolve variants caused by minor title differences or missing DOIs. This procedure ensured that each publication was represented once in the dataset, forming an accurate basis for subsequent screening and eligibility assessment.

These searches were applied to titles, abstracts, and author keywords, and restricted to English-language publications between 2010 and 2025. The initial search returned 353 unique documents, representing studies on optimisation methods applied to building energy retrofit. During the Screening stage, a preliminary evaluation was undertaken based on titles, keywords, abstracts, and publication years, resulting in 285 documents passing the initial assessment. A more detailed review, considering methodology and conclusions, reduced the number of relevant papers down to 187. In the Eligibility stage, the focus was restricted to studies explicitly addressing multi-objective optimisation in building retrofits, leaving 113 eligible documents. Finally, in the Inclusion stage, 45 studies were selected for detailed review. These sources addressed retrofit strategies incorporating multi-objective optimisation, Non-dominated Sorting Genetic Algorithm II (NSGA-II), or Pareto-based genetic algorithms. The selected papers were further categorised by climate type, comprising 38 studies from non-hot-humid climates and seven from hot-humid climates. This final selection underscores the specific research focus on multi-objective optimisation across diverse climatic contexts. The corresponding search terms are consolidated under the stated keywords and presented in

Table 2. Related terms are grouped under the stated keyword.

Keyword	Terms Grouped with the Keyword
Buildings	building; building energy efficiency; building stock
Energy efficiency	energy saving; energy performance; energy savings; building energy performance
Retrofit	building retrofit; building renovation; building energy retrofit; building retrofitting; energy retrofit; retrofitting; existing buildings; existing building; renovation energy retrofits; refurbishment; energy-efficient retrofit; energy retrofitting; deep energy retrofit; retrofits; building refurbishment; energy renovation; retrofit measures; green retrofit; green retrofitting; retrofit scenarios; retrofitting strategies; deep renovation; deep retrofit; energy efficiency retrofit; sustainable building renovation; sustainable building upgrade; zero energy building renovation; building energy renovation; building envelope retrofits; cost-optimal retrofit; façade retrofit; retrofitting measures; social housing retrofit; sustainable retrofit; thermal retrofit; renovation strategies; residential building retrofit; retrofit interventions
Optimisation	Multi-objective optimisation; multi-variable optimisation; multi-criteria optimisation; multiple-objective decision-making; multi-criteria decision-making; Multi-Objective Particle Swarm Optimisation (MOPSO). Pareto optimisation; Pareto front; Pareto optimal solutions, objective functions; decision variables; constraints, Optimal trade-off; optimal retrofit solutions/options/decision/measures
Genetic algorithm (GA)	Multi-criterion GA, Pareto GA, multi-objective GA; two-objective GA; Non-dominated Sorting Genetic Algorithm II (NSGA-II)
Thermal comfort	Adaptive thermal comfort; indoor thermal comfort; outdoor thermal comfort
Hot-humid climates	tropics; tropical climates, hot and humid climate; hot and humid climates; hot/warm and humid climates

.

The Eligibility stage applied explicit inclusion and exclusion criteria. The inclusion criteria are as follows: (1) studies applying multi-objective optimisation (MOO) or genetic algorithms to building retrofit; (2) studies integrating energy, cost, comfort, or environmental performance metrics; (3) simulation-based retrofit analyses using tools such as EnergyPlus, TRNSYS, DesignBuilder, IES-VE, OpenStudio, or equivalent; (4) MOO combined with Pareto-based methods (e.g., NSGA-II, aNSGA-II, MOPSO). The exclusion criteria are as follows: (1) studies focused exclusively on new building design; (2) purely descriptive retrofit studies without quantitative optimisation; (3) non-English publications; (4) papers using decision-making methods alone (e.g., pure AHP/TOPSIS) without simulation-based optimisation; (5) studies addressing general energy efficiency without retrofit-specific interventions. This final selection reflects the core research focus of the review: evaluating the role of multi-objective optimisation in retrofit decision-making across climates, with specific emphasis on hot-humid regions.

3. Results

Alongside a systematic review, a bibliometric analysis was conducted to examine thematic structures, co-occurrence patterns, and research trends within the wider optimisation and retrofit studies. This section reports on data analysis conducted using the bibliometric mapping software VOSviewer version 1.6.20, https://www.vosviewer.com/ (accessed 15 March 2025), which enables the generation of clear and informative visualisations of bibliometric networks. As a freely available tool, VOSviewer is particularly effective for representing large datasets in an accessible and interpretable format. In this study, bibliographic records were exported in RIS format, and a keyword co-occurrence analysis was performed to generate the corresponding bibliometric maps. Figure 2 presents a network visualisation produced from a dataset of 353 research articles on building retrofit optimisation. The map illustrates the relationships between terms derived from titles, abstracts, and author keywords, which are organised into clusters based on their frequency of co-occurrence. Four clusters were identified, each represented by a distinct colour (red, green, yellow, and blue). The red cluster includes terms such as research, retrofit, problem, challenge, and investment, reflecting a thematic focus on research frameworks and implementation challenges in retrofitting. The green cluster comprises terms including climate, technique, insulation, efficiency, and multi-objective optimisation, highlighting themes of thermal performance and optimisation strategies. The yellow cluster contains terms such as energy model, cost, sustainability, and optimal solution, suggesting an emphasis on sustainability assessment and energy modelling. Finally, the blue cluster features terms including decision-making, carbon emissions, framework, and stakeholder, pointing to themes of decision-making, environmental impacts, and strategic planning.

Node size reflects the frequency of occurrence of terms, with larger nodes (e.g., retrofit, efficiency, and decision-making) denoting more prominent research topics. Links between nodes represent co-occurrences, with line thickness indicates the strength of association. Dense connections, particularly in the central regions of the map, reveal a high degree of interrelatedness between terms, underscoring the multidisciplinary nature of building retrofit research. The overlap across clusters suggests that domains such as energy modelling, thermal efficiency, decision-making, and sustainability are strongly interconnected. Notably, terms including multi-objective optimisation, efficiency, and investment indicate both established and emerging research themes. On this basis, multi-objective optimisation is identified as the central focus of this review. Figure 3 further illustrates the co-occurrence links specifically associated with the keyword ‘multi-objective optimisation’.

The primary publication sources on optimisation in building retrofitting are presented in

Table 3. Sources of the primary studies covering the optimisation of building retrofit from 2010 to 2025.

	Sources Title	Count
1	Energy and Buildings	47
2	Applied Energy	25
3	Energies	21
4	Sustainability (Switzerland)	18
5	Journal of Building Engineering	16
6	Energy	15
7	Sustainable Cities and Society	11
8	Building and Environment	11
9	Renewable and Sustainable Energy Reviews	7
10	Journal of Cleaner Production	8
11	Energy Procedia	6
12	E3S Web of Conferences	5
13	Buildings	5

. Energy and Buildings, Applied Energy, and Energies emerged as the leading journals in this domain, contributing 47, 25, and 21 publications, respectively.

As illustrated in Figure 4, the dataset comprised 248 journal articles, 61 conference papers, 29 review papers, and 15 book chapters. Collectively, these outputs address a wide spectrum of themes encompassing energy efficiency, sustainability, and engineering. The findings indicate that most of the research in building retrofit optimisation is disseminated through journal articles, with particularly strong contributions from high-impact journals dedicated to energy and sustainability.

The review of building types represented in the selected literature on multi-objective optimisation indicated a predominant focus on residential buildings. As illustrated in Figure 5, residential buildings constitute 56% of the sample, accounting for 25 of the 45 case studies, reflecting a strong research emphasis on optimising household energy performance. Educational buildings comprised 22% (ten cases), highlighting efforts to improve energy efficiency in schools and universities. Office buildings represented 11% (five cases), while other categories, including hospitals, museums, public buildings, and virtual buildings, were underrepresented, each accounting for only 2–5% of the total. This distribution demonstrates that research on optimisation in building retrofits has been concentrated primarily on the residential and educational sectors, with limited attention given to other building types. Broadening future investigations to encompass a wider range of building categories would facilitate more comprehensive insights into the application of multi-objective optimisation across the built environment.

The geographic distribution of case studies on multi-objective optimisation in building retrofitting demonstrates a diverse yet uneven international representation, as illustrated in Figure 6. A total of 45 studies were identified, with the majority concentrated in a small number of countries. Italy accounted for the largest share, with nine studies, followed by Iran with seven and Portugal with four. France, Malaysia, Spain, and China each contributed between two and three studies, while single case studies were reported from Germany, Turkey, Argentina, Canada, Egypt, Japan, Korea, the United Arab Emirates, the United Kingdom, the United States, Singapore, Brazil, the Philippines, Indonesia, and Panama. Overall, the distribution reveals a strong concentration in Europe and Asia, with limited contributions from Africa and Oceania. The underrepresentation of many regions highlights a significant gap in the global research landscape, indicating the need for wider international engagement and a broader range of case studies. This imbalance also raises concerns regarding the generalisability of existing findings, emphasising the importance of incorporating more diverse geographical and climatic contexts in future investigations on building retrofitting and energy efficiency optimisation.

3.1. Multi-Objective Optimisation

Multi-objective optimisation (MOO), also referred to as multi-criteria optimisation, addresses problems characterised by the presence of multiple, often conflicting, objectives. In contrast, traditional single-objective optimisation seeks to identify the optimal solution for a single performance function. A review of the literature [14] indicates that 82% of the examined studies adopted multi-criteria approaches, while 18% employed single-criteria formulations. In terms of methodology, 55% of the reviewed documents utilised optimisation techniques and 45% applied scenario analysis. Among optimisation methods, approximately 41% employed genetic algorithms, with the Non-dominated Sorting Genetic Algorithm II (NSGA-II) accounting for 76% of these applications and the Single-Objective Genetic Algorithm (SOGA) for 24%. Within the context of building retrofit optimisation, NSGA-II is generally favoured due to its capacity to manage the complexity of multiple objectives and to generate Pareto-optimal solutions. This enables decision-makers to evaluate trade-offs across criteria such as energy efficiency, occupant comfort, cost, and environmental performance. SOGA, by contrast, may be suitable in cases where retrofit projects are driven by a single, dominant objective, and trade-offs are of lesser importance [14].

In practical retrofit applications, decision-making rarely involves a single objective; rather, multiple and often competing criteria must be addressed simultaneously. Multi-objective optimisation has therefore become an important tool for guiding retrofit strategies, offering a structured means of balancing conflicting priorities [16]. Such methods facilitate the integrated evaluation of diverse performance indicators, including energy savings, investment costs, payback periods, and environmental impacts, thereby supporting more holistic and sustainable retrofit decision-making [17]. Key concepts and components of multi-objective optimisation involve the following, as outlined below:

3.1.1. Objective Functions

In multi-objective optimisation, two or more objectives are typically considered simultaneously, reflecting different dimensions of a problem that cannot be meaningfully combined into a single criterion [18]. Objective functions define the goals of the optimisation process, and in building retrofitting, they are often categorised into environmental, economic, and social criteria. As summarised in

Table 4. Retrofit optimisation objectives and criteria.

Objectives	Criteria
	Environmental	Economic	Social
	Minimise energy consumption	Payback period	Thermal comfort
	Minimise energy demand	Initial investment cost	Visual comfort
	Maximise energy savings	Cost-saving	Indoor environmental quality (IEQ)
	Minimise damage to ecosystem quality	Net present value (NPV)	Weighted discomfort time (WDT)
	Minimise depletion of natural resources	Energy cost (EC)	Heritage conservation
	Maximise energy potential	Maintenance cost (MC)	Acoustic comfort
	Minimise OTTV	Replacement cost (RC)
	Minimise CO2 emissions	Residual value (RV)
	Minimise global warming potential	Government incentives
	Minimise embodied emissions/energy	Life-cycle cost (LCC)

, environmental objectives commonly include minimising energy consumption and demand, maximising energy savings, and reducing CO₂ emissions. Economic objectives may address metrics such as payback period, initial investment cost, net present value, and life-cycle cost. Social objectives often focus on enhancing thermal comfort, improving indoor environmental quality, and reducing discomfort hours. Balancing these objectives is central to achieving effective retrofit outcomes.

3.1.2. Pareto Front

The Pareto front (or frontier) represents the set of solutions for which no alternative performs better across all objectives; each solution is therefore considered non-dominated. The identification of the Pareto front constitutes the primary aim of multi-objective optimisation, as it depicts the trade-offs between competing objectives [19]. Pareto-optimal solutions, which lie on this frontier, embody the best achievable compromises, since improvement in one objective inevitably entails a reduction in performance in another [19,20]. Pareto-based optimisation methods explicitly target these solutions, with algorithms such as the Non-dominated Sorting Genetic Algorithm II (NSGA-II) and Multi-Objective Particle Swarm Optimisation (MOPSO) widely applied in building retrofit studies. These methods operate by generating feasible design or retrofit options and identifying those that are non-dominated, thereby forming the Pareto front as the representation of optimal trade-offs. Among such approaches, population-based genetic algorithms (GA) have consistently emerged as the leading techniques for solving optimisation problems in the building sector [19,20].

3.1.3. Decision Variables

Decision variables represent the parameters that can be adjusted within the optimisation process to identify optimal solutions. In the context of building retrofitting, these variables correspond directly to retrofit measures or design choices, which are systematically modified during optimisation to achieve the desired objectives. Examples include insulation thickness (m), window U-value (W/m²K), HVAC setpoints (e.g., temperature and relative humidity), lighting control strategies (such as occupancy sensors and dimming systems), renewable energy system capacities, as well as the sequencing and phasing of retrofit measures.

3.1.4. Constraints

Constraints define the boundaries within which decision variables may operate, ensuring that generated solutions are both feasible and practically implementable [19]. These may include economic constraints, such as budgetary limits on total retrofit expenditure; physical constraints, such as maximum allowable insulation thickness; comfort-related constraints, including maintaining indoor temperatures above specified thresholds; and regulatory constraints, such as adherence to building codes, minimum fresh air requirements, or compliance with energy performance standards. By restricting solutions to those that respect real-world limitations, constraints ensure the relevance and applicability of optimisation outcomes.

Multi-objective optimisation has been widely applied across many disciplines, including engineering, finance, and environmental management, as well as complex decision-making, where multiple conflicting objectives must be addressed simultaneously. Its application in building retrofitting is particularly valuable as it allows for the systematic balancing of competing priorities such as performance, cost, comfort, and sustainability. Studies consistently highlight its capacity to enhance building energy efficiency and environmental performance. Attia et al., [21] concluded that multi-objective optimisation is one of the most robust forms of optimisation because it generates sets of solutions from trade-offs between multiple design objectives that contradict each other.

3.2. Existing Review in Multi-Objective Optimisation Building Energy Retrofit

The multi-objective optimisation of building energy retrofits involves the integration of advanced modelling techniques, optimisation algorithms, and decision-making frameworks to achieve improvements in energy efficiency, cost-effectiveness, and environmental performance. For example, an early literature review by Evins [22] examined computational optimisation techniques applied to building envelopes, systems, and energy generation between 1990 and 2012. Similarly, through interviews with Building Performance Optimisation (BPO) experts, Attia et al. [21] identified critical gaps in the integration of BPO tools within retrofit practice. Together, these studies have offered comprehensive perspectives on multi-objective optimisation approaches, supporting retrofit planning and decision-making processes by enabling stakeholders to balance competing objectives and constraints in pursuit of energy efficiency, cost-effectiveness, and sustainability.

Subsequent research has expanded the scope of inquiry to include multi-modal optimisation problems, the comparative performance of optimisation algorithms, applications of surrogate modelling, optimisation under uncertainty, and challenges in operationalising optimisation techniques within real-world building design and retrofit contexts [23]. Jafari and Valentin [24] further reviewed the selection of optimisation objectives for decision-making in retrofit projects, emphasising the economic, environmental, and social benefits of retrofitting, as well as the use of decision matrices in case-study applications.

Systematic reviews of genetic algorithm-based multi-objective optimisation highlight the significant potential of these techniques for developing retrofit strategies that address multiple targets simultaneously, including operational energy demand, retrofit costs, and environmental impacts. Costa-Carrapiço et al. [19] underscored the importance of explicitly considering trade-offs among multiple objectives to achieve optimal retrofit outcomes. Similarly, Hashempour et al. [14] advocated for the development of decision-support tools to enhance the energy performance of existing buildings, while also noting that most empirical studies remain concentrated in developed countries.

It is evident from the review of recent MOO energy retrofit research that little work has been done on this topic for buildings in countries that experience hot-humid climates. One aim of this review was to highlight this and to encourage more MOO energy retrofit work related to these countries.

3.3. Previous Study of Multi-Objective Optimisation Building Energy Retrofit

Several case studies have applied genetic algorithms and other heuristic methods within multi-objective optimisation (MOO) frameworks to support energy retrofit decision-making. Collectively, these studies demonstrate the potential of advanced optimisation techniques to enhance energy efficiency, cost-effectiveness, and sustainability in building retrofits. Key insights from the literature are summarised in Table 5.

Multiple objective optimisation models were developed for building energy efficiency investment decisions to maximise energy savings and minimise the payback period. This model assists decision-makers in making optimal choices when investing in energy-efficient building retrofitting projects [17]. Subsequent research broadened methodological approaches through the application of Multi-objective Particle Swarm Optimisation (MOPSO) and Multi-objective Ant Colony Optimisation (MACO) to evaluate thermal comfort, operational energy demand, and life-cycle energy use, highlighting the versatility of optimisation techniques in performance analysis [23]. Asadi et al. [25] further demonstrated the use of Genetic Algorithms (GA) in combination with Artificial Neural Networks (ANN) to optimise energy consumption, retrofit costs, and thermal discomfort hours, reinforcing their value in retrofit decision-making. Similarly, Penna et al. [26] applied NSGA-II coupled with dynamic simulation tools to examine the role of reference-building characteristics in defining optimal retrofit strategies. Their findings suggested that while conventional energy efficiency measures (EEMs) could achieve near-zero energy targets within economic constraints, they may compromise indoor thermal comfort.

More recent studies have advanced the application of MOO in various retrofit contexts. Merlet et al. [27] demonstrated the effectiveness of MOO in generating multiple optimal retrofit scenarios for building stock upgrades, enabling experts to select contextually appropriate solutions. Pilechiha et al. [28] applied a multi-objective framework to office window design, balancing view quality, daylight access, and energy efficiency, illustrating the complexity of optimisation in commercial buildings. In the context of historic building retrofits, Qu et al. [29], emphasised the importance of balancing insulation, airtightness, and shading measures to improve thermal comfort alongside energy performance.

Algorithmic advancements have also been notable. Rosso et al. [30] employed an archive-based NSGA-II (aNSGA-II) for residential retrofits in Rome, significantly reducing computational time while identifying solutions capable of reducing annual energy demand by 49.2%, energy costs by 48.8%, and CO₂ emissions by 45.2%, with nearly 60% lower investment costs compared to other criterion-optimal solutions. Similarly, Lu et al. [31] highlighted the importance of occupant-centred retrofits, showing that measures such as occupancy-based lighting sensors, higher temperature setpoints, and reduced plug loads could outperform more technology-driven strategies like chiller replacement or green roof installation.

Shao et al. [32] examined high-rise residential buildings in Northwestern China, applying the SPEA-2 algorithm across four objectives: life-cycle cost, energy consumption, carbon emissions, and thermal comfort, and eleven design variables. Their results demonstrated that optimisation outcomes are sensitive to climatic factors, such as solar radiation and to the choice of objectives themselves. Dehghan and Porras Amores [33] reported that MOO-based retrofit strategies for residential buildings could reduce primary energy consumption and greenhouse gas emissions by 60%, thermal discomfort by 65%, and visual discomfort by 83%, while maintaining indoor air quality within ASHRAE standards. Nonetheless, they stressed the need for careful consideration of trade-offs in design decisions, particularly under climate change pressures.

Despite these advances, the majority of MOO retrofit studies have been conducted in subtropical climates. Tavakolan et al. [34] emphasised the need for research across different building types and climate zones, particularly hot-humid regions. Expanding on this point, Aruta et al. [35] argued that systematic climate-specific analyses of building characteristics are essential for identifying the most effective retrofit measures.

3.4. Characteristics of Effective Retrofit Strategies in Hot-Humid Climates

Developing retrofit strategies for buildings in hot-humid climates requires an integrated approach that combines passive and active design solutions to reduce cooling loads while maintaining occupant comfort. Because temperature and humidity remain high throughout the year, the dominant challenge is limiting solar heat gain and managing latent cooling energy rather than achieving thermal insulation against the cold. As several authors have emphasised, climate-specific retrofitting is essential, since measures effective in temperate or arid regions, such as heavy insulation or heat-recovery ventilation, can perform poorly or even increase discomfort in hot-humid environments [36,37]. Consequently, retrofit solutions must target solar control, natural ventilation, and efficient cooling, balancing energy performance with indoor air quality and thermal comfort.

3.4.1. Passive Strategies

Passive measures form the first line of defence against excessive heat gains. Enhancing the building envelope through reflective roof coatings, insulated walls, and airtight construction can substantially reduce cooling energy demand. Multiple studies have shown that improving envelope performance can lower annual cooling loads by 20–40% in tropical office buildings [38,39]. Roof insulation is particularly influential because of the high incident solar radiation on horizontal surfaces, while the use of high-albedo materials limits heat absorption.

3.4.2. Glazing and Shading

Glazing and shading systems are equally critical. Research by Ayodele et al. [40] demonstrated that electrochromic glazing combined with external shading can reduce heat gain by up to 59%, whereas unshaded glass significantly increases discomfort. Similarly, [41] highlighted that low-emissivity glazing and heat-capacitive wall materials help stabilise indoor temperatures. The effectiveness of shading devices, horizontal louvres, fins, overhangs, and light shelves has been verified across multiple studies in Southeast Asia [42,43]. These elements intercept direct solar radiation and reduce glare while maintaining daylight, offering a low-cost and low-maintenance strategy suitable for both new and existing façades.

3.4.3. Active Strategies

Where passive measures alone cannot maintain acceptable comfort, active systems, such as HVAC and lighting retrofits, become necessary. Optimising air-conditioning systems by upgrading to variable-speed compressors, high-efficiency chillers, and intelligent control systems can cut cooling energy by 15–30% [7,44]. Equally important is recalibrating thermostat set-points and zoning systems to align with adaptive comfort standards, which reduces over-cooling common in tropical offices. Integrating energy-efficient lighting systems, such as LEDs with occupancy or daylight sensors, further diminishes internal heat gains and electricity use. Studies consistently show that lighting retrofits can yield 20–40% electricity savings while improving illumination quality [7].

3.4.4. Renewable and Integrated Systems

Beyond passive and active measures, renewable-energy integration plays an expanding role in deep retrofits. Photovoltaic (PV) systems and hybrid solar-cooling technologies can offset part of the building’s electricity consumption and contribute to long-term carbon neutrality [4,39]. Energy and CFD simulations in heritage buildings have also demonstrated the potential of night ventilation strategies to pre-cool thermal mass and maintain acceptable humidity without mechanical cooling [45]. These approaches highlight the value of hybrid solutions combining renewable energy generation with passive cooling.

3.4.5. Energy-Management and Building-Automation Systems

At the operational level, energy-management and building-automation systems provide the analytical backbone for optimising performance. Real-time monitoring and control of HVAC, lighting, and plug loads enabled facility managers to identify inefficiencies, adjust schedules, and verify post-retrofit savings [8,9]. Energy auditing and performance assessment stress how conducting thorough energy audits and performance assessments can help identify areas of improvement and prioritise retrofit measures [4,12,13]. The increasing availability of Internet-of-Things (IoT) devices has further enhanced the practicality of continuous commissioning and user feedback loops, which are essential in climates where occupancy patterns and comfort preferences strongly influence performance.

3.4.6. Occupant-Behaviour Interventions

Occupant-behaviour interventions also represent a cost-effective opportunity. Adjusting thermostat settings, improving user awareness of adaptive comfort, and employing occupancy-based controls can yield immediate energy savings without capital expenditure. Studies [31,46] have demonstrated that behaviour-based measures can rival or even surpass some technological retrofits when measured in terms of cost-effectiveness and speed of implementation.

3.5. Multi-Objective Optimisation Retrofit Studies in Hot-Humid Climate

Optimising energy retrofits improves building performance, reduces energy consumption, and enhances occupant comfort across diverse climatic conditions. In hot-humid or tropical regions, characterised by year-round hot and rainy seasons, buildings, particularly multi-storey and large-scale developments, rely heavily on air conditioning to maintain acceptable indoor environments. Despite this pressing energy demand, relatively few studies have examined optimisation methods, techniques, and parameters specifically tailored to energy retrofits in such tropical climates. Notable multi-objective optimisation (MOO) studies have been conducted in Singapore, Brazil, the Philippines, Malaysia, Indonesia, and Panama.

Occupant-oriented strategies have been shown to outperform purely technological retrofits. For example, Lu et al. [31] demonstrated that measures such as occupancy-based lighting sensors, higher thermostat setpoints, and plug-load reductions yielded greater energy savings than technological upgrades such as chiller replacement or green roof installation. In Brazil, Naves et al. [47] applied MOO using TRNSYS 18 Simulation Studio and GenOpt to a seven-storey residential building, optimising life-cycle cost (LCC) and payback period while considering photovoltaic (PV) generation, grid consumption, chillers, and thermal storage. Seghier et al. [48] proposed a BIM-integrated optimisation methodology for an office building in Malaysia to customise NSGA-II, with objectives focused on minimising retrofit costs and reducing the Overall Thermal Transfer Value (OTTV), that is, the average heat gain into the building through the building envelope.

Further contributions include an Indonesian case study where a MOO framework optimised cooling energy demand and discomfort hours in office buildings by varying window-to-wall ratios, glazing type, blinds, and shading systems [49]. In Panama, Gonzalez et al. [50] applied NSGA-II and the dynamic thermal simulation software DesignBuilder to a low-rise apartment building, minimising life-cycle cost and net primary energy through strategies such as glazing, local shading, cooling schedules, external wall construction, site orientation, and PV integration.

Complementary approaches using multi-criteria decision-making (MCDM) have also been applied. In Malaysia, Balasbaneh et al. [51] assessed operational energy, global warming potential, embodied energy, and retrofit costs, identifying double-glazing as the optimal window replacement option. In the Philippines, Ongpeng et al. [52] employed an MCDM approach combining Analytic Hierarchy Process (AHP) and the VIKOR method to optimise retrofit scenarios for an educational building, balancing investment cost, payback period, and environmental performance through improvements to the envelope, mechanical and electrical systems, and on-site energy generation. Collectively, these studies highlight the applicability of multi-objective optimisation and MCDM frameworks in the context of hot-humid climates, addressing environmental, economic, and social objectives. However, the limited number of case studies underscores the need for broader investigations into retrofit optimisation tailored to the distinctive challenges of tropical climates. Table 5 summarises the selected studies conducted in these regions.

Table 5. Selected previous studies of multi-objective optimisation retrofit in hot-humid climates (in chronological order).

Authors	Year	Case studies	Location	Methods	Software	Decision Matrix Objectives			Energy Efficiency Measures (EEMS)
						Economic	Environmental	Social
Lu et al. [31]	2021	School of Design and Environment, National University of Singapore	Singapore	Calibration; cost–benefit analysis decision-making for building energy efficiency retrofit	IES-VE Appro profiles ApacheSim Vistapro	Investment Cost (IC); Net Present Value (NPV); Benefit–Cost Ratio (BCR); life cycle cost (LCC)	Life Cycle Sustainability Evaluation (LCSE)	-	HVAC; lighting equipment; façade (windows, wall, roof, openings)
Naves et al. [47]	2021	Residential building, seven storeys	Brazil, at Rio de Janeiro (23° S Latitude)	Multi-objective optimisation	TRNSYS 18 Simulation Studio, GenOpt (Generic Optimisation Program)	life-cycle cost (LCC) and payback period	PV energy (production, consumption, and surplus), grid energy consumption	-	PV; grid; chiller; thermal energy storage
Ongpeng et al. [52]	2022	National Engineering Center, University of the Philippines, four floors	Philippines	Weighing method; AHP compromise ranking method; VIKOR	DesignBuilder	Investment Cost (IC) Payback Period	Damage to Ecosystem Quality Depletion of Natural Resources Energy Potential	Damage to Human Health	Building envelope; mechanical system; electrical system; on-site generation
Seghier et al. [48]	2022	Office Building University Teknologi Malaysia	Malaysia	NSGA-II	Autodesk Revit, Visual Scripting: Dynamo, NSGA II: MATLAB	Minimise retrofit cost	Minimise OTTV	-	Walls; windows
Balasbaneh et al. [51]	2022	The room’s windows	Malaysia	Multi-criteria decision-making (MCDM)	EnergyPlus	Cost of each alternative	Operation energy usage; global warming potential (GWP) emission; embodied energy	-	Windows
Ardiani et al. [49]	2024	Office Building	Jakarta, Indonesia	NSGA-II	DesignBuilder	-	Minimising cooling energy	Reducing discomfort hours	Window-to-wall ratio, glazing type, window blind type, and shading type
Gonzales et al. [50]	2024	Low-rise apartment building	Panama	NSGA-II	DesignBuilder	Minimise life-cycle cost	Minimise net primary energy	-	Glazing type; local shading type; cooling operation schedule; external wall construction; site orientation; PV application

Table 5. Selected previous studies of multi-objective optimisation retrofit in hot-humid climates (in chronological order).

Authors	Year	Case studies	Location	Methods	Software	Decision Matrix Objectives			Energy Efficiency Measures (EEMS)
						Economic	Environmental	Social
Lu et al. [31]	2021	School of Design and Environment, National University of Singapore	Singapore	Calibration; cost–benefit analysis decision-making for building energy efficiency retrofit	IES-VE Appro profiles ApacheSim Vistapro	Investment Cost (IC); Net Present Value (NPV); Benefit–Cost Ratio (BCR); life cycle cost (LCC)	Life Cycle Sustainability Evaluation (LCSE)	-	HVAC; lighting equipment; façade (windows, wall, roof, openings)
Naves et al. [47]	2021	Residential building, seven storeys	Brazil, at Rio de Janeiro (23° S Latitude)	Multi-objective optimisation	TRNSYS 18 Simulation Studio, GenOpt (Generic Optimisation Program)	life-cycle cost (LCC) and payback period	PV energy (production, consumption, and surplus), grid energy consumption	-	PV; grid; chiller; thermal energy storage
Ongpeng et al. [52]	2022	National Engineering Center, University of the Philippines, four floors	Philippines	Weighing method; AHP compromise ranking method; VIKOR	DesignBuilder	Investment Cost (IC) Payback Period	Damage to Ecosystem Quality Depletion of Natural Resources Energy Potential	Damage to Human Health	Building envelope; mechanical system; electrical system; on-site generation
Seghier et al. [48]	2022	Office Building University Teknologi Malaysia	Malaysia	NSGA-II	Autodesk Revit, Visual Scripting: Dynamo, NSGA II: MATLAB	Minimise retrofit cost	Minimise OTTV	-	Walls; windows
Balasbaneh et al. [51]	2022	The room’s windows	Malaysia	Multi-criteria decision-making (MCDM)	EnergyPlus	Cost of each alternative	Operation energy usage; global warming potential (GWP) emission; embodied energy	-	Windows
Ardiani et al. [49]	2024	Office Building	Jakarta, Indonesia	NSGA-II	DesignBuilder	-	Minimising cooling energy	Reducing discomfort hours	Window-to-wall ratio, glazing type, window blind type, and shading type
Gonzales et al. [50]	2024	Low-rise apartment building	Panama	NSGA-II	DesignBuilder	Minimise life-cycle cost	Minimise net primary energy	-	Glazing type; local shading type; cooling operation schedule; external wall construction; site orientation; PV application

3.6. Cross-Study Patterns of Retrofit Effectiveness in Hot-Humid Climates

A key contribution of this review lies in advancing a climate-specific synthesis of optimisation-based retrofit strategies for hot-humid regions, an aspect that has received little attention in previous reviews. Earlier optimisation surveys have catalogued algorithms, decision variables, and performance outcomes, but they have not developed a framework tailored to the distinctive thermal, moisture, and operational conditions of hot-humid climates. Drawing on the seven optimisation studies identified within such regions, this review introduces a structured matrix that organises retrofit decision-making around three interrelated dimensions: objectives, measures, and constraints. Typical optimisation objectives in hot-humid contexts include reducing cooling energy use, lowering life-cycle cost, improving OTTV performance, and decreasing thermal discomfort hours. These objectives are evaluated against a set of measures that consistently emerge as influential in tropical climates, such as window-to-wall ratio, glazing solar heat-gain coefficient, shading configuration, cooling setpoint adjustment, operational scheduling, and photovoltaic deployment, while acknowledging constraints that are largely absent in temperate studies, including persistent high humidity, elevated latent loads, reduced night-time cooling potential, and urban heat-island amplification. This framework provides an interpretive lens not found in previous optimisation reviews and allows for climate-specific retrofit strategies to be more coherently understood.

Beyond establishing this conceptual structure, the review offers a synthesis of dominant optimisation themes that recur across hot-humid case studies. Three findings are particularly consistent. First, operational strategies, such as cooling setpoint elevation, plug-load management, or schedule optimisation, frequently outperform technology-only retrofits, demonstrating that behavioural and controls-based interventions can deliver substantial cooling reductions at low cost. Second, envelope modifications in hot-humid regions are driven more by the need to reduce solar heat gain than by increasing insulation thickness; shading devices and low-SHGC glazing reliably yield greater benefits than conventional insulation upgrades. Third, photovoltaic integration, especially when combined with storage, emerges as a robust Pareto-optimal solution in climates with high daytime cooling loads and strong solar resources. These themes collectively articulate a climate-responsive optimisation logic that is largely absent from general retrofit reviews.

To complement this thematic synthesis, the review also develops a taxonomy of tropical retrofit decision variables, classifying the parameters most frequently used in optimisation models for hot-humid buildings. These include window-to-wall ratio, glazing type and SHGC, shading geometry, OTTV-related envelope parameters, photovoltaic capacity, and operational schedules, variables that differ substantially from those that dominate optimisation studies in heating-oriented climates. Based on this taxonomy and on the comparative evidence in the seven hot-humid studies, the review further identifies cross-study patterns in measure effectiveness. High-impact measures with strong consensus include cooling setpoint optimisation, external shading strategies, low-SHGC glazing, and schedule-based reductions in cooling and lighting loads. Medium-consensus measures, such as PV plus storage, show promising performance but are less frequently evaluated. Some measures, including façade greening and green roofs, offer benefits yet appear only in isolated studies and are seldom selected as optimal within multi-objective algorithms. This qualitative meta-synthesis provides the first consolidated evidence ranking retrofit strategies specifically for hot-humid climates.

Taken together, these contributions, an original climate-specific optimisation framework, the identification of dominant optimisation dynamics, a tailored taxonomy of decision variables, and a comparative synthesis of measure effectiveness, represent analytic advances not found in prior reviews of multi-objective retrofit optimisation. They demonstrate how retrofit optimisation operates differently in hot-humid climates and provide a clearer foundation for future methodological and applied research in these regions.

Table 6. Cross-study patterns of retrofit effectiveness in hot-humid climates.

Component	Content	Description/Rationale
OBJECTIVES	Cooling energy reduction, life-cycle cost minimisation, OTTV reduction, reduction in thermal discomfort hours, CO2/primary energy reduction (when included)	Objectives in hot-humid regions strongly emphasise cooling load control and economic performance, reflecting persistent high temperatures and humidity. OTTV is regionally mandated in several tropical countries.
OPTIMISATION MEASURES	Envelope and Solar Control: glazing SHGC/U-value, window-to-wall ratio (WWR), external shading (fins, overhangs, blinds); HVAC and Operations: cooling setpoints (increased by 1–2 °C), occupancy- and schedule-based control, ventilation strategy tuning; Energy Systems: photovoltaics (PV) ± storage, efficient chiller operation.	Measures reflect the climatic need to limit solar heat gain, manage internal gains, and moderate cooling demand. Operational measures often outperform physical upgrades in cost-effectiveness.
CLIMATE-SPECIFIC CONSTRAINTS	High latent loads and humidity control requirements, reduced night-time cooling potential, high external humidity and infiltration sensitivity, solar exposure and glare constraints, urban heat-island. effects, common tropical building archetypes (e.g., large glazing areas).	These constraints shape optimisation search spaces and explain why certain measures—like shading and glazing optimisation—are consistently selected by MOO algorithms.
DOMINANT PATTERNS EMERGING FROM REVIEWED STUDIES	Operational optimisation frequently outperforms technology-only retrofits, solar heat-gain control has greater impact than insulation thickness, PV integration often appears on the Pareto front under high cooling loads, low-SHGC glazing and shading yield high consensus as effective interventions.	These patterns provide generalised design guidance and represent insights not offered in previous retrofit optimisation reviews.
TYPICAL HIGH-IMPACT MEASURES (CONSENSUS ACROSS STUDIES)	Increase cooling setpoints by 1–2 °C, add external shading or improve shading geometry, use low-SHGC glazing, optimise AC and lighting schedules.	Consistently identified as effective in case studies from Singapore, Malaysia, Indonesia, Brazil, Panama, and the Philippines.
MODERATE/CONTEXT-SPECIFIC MEASURES	PV + storage (context-dependent economics), wall/roof insulation (limited effect unless façade heavily sun-exposed), green roofs/façade greening (effective but rarely Pareto-optimal).	Helps readers understand relative performance without overstating numerical results.

provides cross-study patterns of retrofit effectiveness in hot-humid climates.

4. Discussion and Conclusions

The application of multi-objective optimisation offers a practical means of balancing energy efficiency, cost-effectiveness, and environmental performance in building retrofits. Advanced computational techniques, including genetic algorithms and decision-making frameworks, enable the simultaneous evaluation of multiple objectives and constraints, thereby enhancing the decision-making process. Such approaches have been extensively applied in studies addressing life-cycle costs, greenhouse gas emissions, and retrofit strategies, particularly in developed countries. More recently, advanced optimisation tools, such as machine learning models and simulation-based methods, have been increasingly employed, demonstrating their capacity to minimise energy use, thermal discomfort, and operational costs. When extended to both building and community scale applications, these methods yield even greater energy savings. Algorithms such as NSGA-II have proven computationally efficient in managing large-scale simulations, supporting stakeholders in identifying and adopting optimal retrofit solutions that are both sustainable and cost-effective.

Some research gaps have been identified from this review. Most studies have been computer-based, and there are very few examples of buildings that have been monitored before and after retrofit measures have been applied, particularly in hot-humid countries. Although retrofit strategies are climatically sensitive, the impact of climate change on retrofit performance has received little attention. Most studies have focused on one building type (i.e., residential) or even just one building (i.e., a large commercial building). This approach makes it difficult to develop evidence-based retrofit strategies that can be applied to large-scale developments or to more unique construction types, such as heritage buildings.

As listed in Table 7, there are significantly different retrofit approaches adopted in hot-humid climates compared to warm or temperate climates. Research indicates that occupant-focused retrofit strategies, such as occupancy-based sensors, adaptive temperature management, and behavioural interventions, often achieve superior outcomes in terms of energy savings and comfort compared with purely technological upgrades. Within the context of a hot-humid climate, optimisation frameworks and multi-criteria decision-making (MCDM) approaches must prioritise objectives such as reducing cooling energy demand, minimising retrofit costs, and enhancing thermal comfort. Effective measures include optimising window-to-wall ratios, glazing configurations, shading devices, and thermal envelope designs, all of which have demonstrated considerable potential to enhance building performance while addressing the distinctive challenges of tropical environments.

The evidence from this review suggests that while a range of optimisation tools and objectives have been applied to retrofitting in hot-humid climates, common themes include the prioritisation of cooling energy reductions, the integration of renewable energy technologies, and the balancing of economic and environmental considerations. Nevertheless, significant gaps remain in the comprehensive integration of occupant comfort, resilience to climatic extremes, and long-term environmental impacts into optimisation frameworks. Addressing these gaps will be essential for developing holistic, resilient, and sustainable retrofit strategies tailored to tropical and subtropical regions.

Table 7. Selected previous studies of multi-objective optimisation retrofit (in chronological order).

Authors	Year	Case Studies	Location	Methods	Software	Decision Matrix Objectives			Energy Efficiency Measures (EEMS)
						Economic	Environmental	Social
Asadi et al. [53]	2012	Semi-detached house (one family) built in 1945	Portugal	Objective function; Thchebycheff programming formulation change; building data from the Portuguese building thermal code (RCCTE)	Bintprog function in MATLAB	Investment cost	Energy for heating, cooling, and water heating	-	Window type: external wall insulation, roof insulation, and solar collector type
Asadi [54]	2012	Residential building	Portugal	NSGA-II	TRNSYS, GenOpt and a Tchebycheff optimisation technique developed in MATLAB	Retrofit cost	Energy savings	Thermal comfort	External wall insulation materials, roof insulation materials, window type, and solar collector type.
Asadi et al. [25]	2014	School building	Coimbra, Portugal	Genetic Algorithm and Artificial Neural Network	TRNSYS, MATLAB	Retrofit cost	Energy consumption	Thermal discomfort hours	Building envelope parameters (including external wall and roof insulation materials, and window type) and the HVAC system type.
Shao et al. [55]	2014	Three-storey office building	Aachen, Germany	Non-dominated Sorting Genetic Algorithm-II (NSGA-II)	House of Quality	Initial investment cost (IC)	Annual operational energy, annual emission GWP, annual energy consumption, envelope air leakage, and climate	Indoor air quality; thermal comfort	Wall insulation, roof insulation, floor insulation, renewable insulation materials, improve building tightness, glazing insulation, advanced envelope technologies, heating/cooling system, building automation system, and photovoltaics
Penna et al. [26]	2015	12 sets of residential buildings	Italy	Genetic Algorithm-II (NSGA-II)	MATLAB TRNSYS 17	Investment cost (IC), net present value (NPV), energy cost (EC), maintenance cost (MC), replacement cost (RC), residual value (RV), and government incentives	Energy performance heating	Weighted discomfort time (WDT)	Wall insulation, roof insulation, floor insulation, glazing system, heating generator, and mechanical ventilation with heat recovery
Carli et al. [56]	2015	Public building	Bari, Italy	Pareto frontier	MATLAB with the Global Optimization Toolbox	the cost and its payoff	Reduction in electrical energy consumption, reduction in methane consumption, reduction in water consumption	Increase in occupants’ internal comfort	Energy efficiency, sustainability, and thermal comfort
Solmaz et al. [57]	2016	Public school building	Izmir, Turkey	Sensitivity analysis; Multi-objective optimisation	Sketch-up Open Studio; Simlab; MATLAB; GenOpt	Net present value (NPV)	Heating and cooling savings	-	Exterior walls, windows, shading, ground floor, and roof
Almeida [58]	2016	School buildings	Portugal	Multi-objective optimisation	DesignBuilder and ANN (Artificial Neural Networks)	Life-cycle cost (LCC)	Energy efficiency	Occupants’ thermal comfort	External wall U values, roof U values, window U values, the total solar energy transmittance of windows, and air change rate (ACR)
Camporeale et al. [59]	2017	Housing blocks	Seville	Genetic algorithm and Pareto rank	Rhinoceros, Grasshopper, and Galapagos; EnergyPlus through DIVA	Net present value (NPV)	Heating and cooling demand	-	Building shape, envelope U values, and window-to-wall ratio (WWR)
Kim et al. [60]	2017	General hospital building	Korea	Multi-objective optimisation	ILOG CPLEX STUDIO 6.0	Retrofit cost	Energy savings	-	Sets of external wall windows and insulation, roof insulation, and solar energy collectors
Ascione [61]	2017	Educational building	Benevento, South Italy	NSGA-II smart exhaustive sampling; Cost-optimal analysis	Energy Plus and MATLAB	Investment cost, global cost	Minimise thermal energy demand for space heating (TEDh) and space cooling (TEDc); annual percentage of discomfort hours (DH)	Thermal comfort	Thermal envelope, HVAC systems and equipment, and renewable energy sources
Jafari and Valentin [24]	2018	Ranch-style home	Albuquerque, NM	A genetic algorithm optimisation	eQuest (Quick Energy Simulation Tool)	Initial investment cost (IC), energy consumption cost (EC), maintenance and replacement cost (MR), property tax (TX), and resale value benefit (RV)	-	-
Bandera et al. [62]	2018	University building	Spain	NSGA-II	Energy Plus jEPlus	-	Minimum energy demand required for heating and cooling, minimise the exergy required, and maximise the energy available	-	Façade, roof, roof skylight, and windows
Bonamente et al. [63]	2018	Fire station	Cuneo, Italy	NSGA-II	Termolog EpiX8	Minimise discounted cost	Minimise thermal energy consumption, minimise electric energy consumption, and minimise GHG emissions	Maximise comfort level	Building insulation, heat generator, thermal distribution system, terminal units, heating control system, electrical systems, solar thermal collectors, and photovoltaic plant
Bosco et al. [64]	2018	Office building	Rome, Italy	NSGA-II	IDA-ICE and MOBO	Total investment cost	Annual total energy consumption	Annual discomfort hours	Four possible wall insulations, five roof insulations (in three possible thicknesses), and five possible windows, with different set-points for the heating curve
Rosso et al. [30]	2020	Residential building, three floors	Rome, Italy	Active-archive Non-dominated Sorting Genetic Algorithm-II (aNSGA-II)	EnergyPlus; Python version 3.6.8	Investment cost (IC) and energy cost (EC)	Energy demand and CO2 emissions	-	Glazing system, opaque vertical envelope insulation system, opaque horizontal envelope insulation system, opaque envelope finishing, layer optics characteristics, solar shading, sun space, closing the courtyard, closing the ground floor, solar thermal collectors, photovoltaic panels, and tilt angle of the thermal and photovoltaic panels
Pilechiha et al. [28]	2020	The office building’s windows	Tehran, Iran	Pareto Frontier and a weighted sum	Daylighting: Grasshopper plug-ins: Ladybug and Honeybee Energy simulation: Energy Plus and Open Studio	-	Minimise energy consumption and maximise the daylight	Visual comfort (absence of glare)	Wall materials, glazing systems, and window size
Rogeau et al. [65]	2020	Virtual building stocks	Rhone district in France	Single objective optimisation	R and RStudio version 3.5.3, and solved with the ILOG IBM CPLEX solver	Net present cost (NPC) and profitability	Energy demand and reduction in GHG emission	-	Building envelope and building heating systems
Aghamolaei et al. [66]	2020	Typical dwellings	Yazd, Iran	Parametric Sensitivity Analysis (PSA) multi-objective optimisation NSGA II	JEPlus and JEPlus-EA	-	Greenhouse gas emissions (GHG)	Improving indoor thermal comfort	-
Amani [67]	2020	Residential apartment building	Tehran, Iran	Multi-objective optimisation NSGA-II Euclidean distance method	DesignBuilder SimaPro	-	Minimise annual energy consumption and minimise global warming potential	-	Insulation system of the external wall
Amiri [68]	2020	A three-story residential building	Montreal, Canada	Bi-objective optimisation	Analytic hierarchy process (ANP) for multiple criteria decision-making (MCDM)	Minimise cost	Maximise utility	-	Passive energy technologies (e.g., window improvement, external wall insulation, internal wall insulation, vapour barrier, and weather barrier)
Ciardiello et al. [69]	2020	Residential apartment block	Rome, Italy	Active-archive non-dominated sorting genetic algorithm (aNSGA-II)	SketchUp, Open Studio, Energy Plus	Annual energy cost and investment cost	Total energy demand and CO2 emissions	-	Insulation on the opaque envelope and transmittance of the glazing system, the optic characteristics of the finishing layer of the external envelope, the presence or absence of brise-soleil and sunspaces on the loggias and in the greenhouse in the courtyard, and photovoltaic and solar thermal panels.
Chang et al. [70]	2020	Two apartments and two wooden houses	North Sumida, Tokyo, Japan	Multi-objective optimisation under uncertainties	Grasshopper in Rhinoceros 3D, Honeybee plugin, MATLAB	Payback period	CO2 emission and energy balance	Discomfort hours	Building envelope design
Qu et al. [29]	2021	Late nineteenth-century Victorian house	United Kingdom	NSGA-II	Energy Plus	Initial investment cost and payback period	Energy savings and annual energy consumption	-	Glazing system, airtightness, and thermal insulation
Hong et al. [71]	2021	Low-rise office building	Shanghai, China	NSGA-II, Linear Programming Techniques for Multidimensional Analysis of Preference (LINMAP)	Rhinoceros, Grasshopper plugins, namely Ladybug and honeybee, Octopus	-	Minimising energy demand	Maximising daylight availability	Thermochromic glazing
Cao et al. [72]	2021	Residential building	Northern Anhui, China	Multi-objective optimisation	DesignBuilder	Retrofit cost	Energy saving	Thermal comfort	External wall insulation layer and setting the roof insulation layer)
Tavakolan et al. [34]	2022	Single-family residence 240m2	Iran	NSGA-II	Building simulation: Energy Plus NSGA-II; MATLAB	Net present value (NPV), investment cost (IC), present value of electricity, natural gas, renewable energy, discounted payback period (DPP)	Primary energy consumption (PEC)	-	Wall Insulations, roof insulations, improving airtightness, window glazing, heating system, cooling system, and photovoltaic panels
Merlet et al. [27]	2022	Social housing stock; three apartment blocks are representative of the whole building’s stock	Paris, France	NSGA-II integration of temporality in optimisation, implementation of sequencing, and implementation of phasing	EnergyPlus in Design Builder DEAP Phyton	Cost	Heating demand and overheating	-	Window properties and wall properties
Swedberg [73]	2022	Multi-family residential projects	Humid Continental Climates (Dfb subtype)	Multi-objective design optimisation (MODO) NSGA-II	EnergyPlus, jEPlus, and jEPlus+EA	-	Minimise heating demand	Overheating-degree-hour (OHDH)	Exterior wall, slab, roof, airtightness, glazing (SHCG), and overhang depth
Aram [74]	2022	Educational building	Tehran, Iran	TOPSIS for retrofit measures	jEPlus, DesignBuilder	Minimise investment cost	Minimise heating and cooling load	-	Building envelopes, insulation, renewable energy, intelligent devices for managing energy consumption, indoor quality, and energy-efficient appliances
Abdelaziz et al. [75]	2023	Villa archetype	Egypt	Multi-objective optimisation	-	Life-cycle cost (LCC)	Life-cycle carbon footprint (LCCF)	-	Insulating external walls and roofs, upgrading existing windows, efficient lighting system, HVAC system, and solar PV cells
Ciardiello et al. [76]	2023	Social housing building	Rome, Italy	Active-archive non-dominated Sorting Genetic Algorithm (aNSGA-II)	Rhinoceros, Grasshopper, EnergyPlus	Investment and operational costs	Minimising energy consumption and CO2 emissions	-	Adding mid and internal thermal insulation to the external cavity walls, adding internal thermal insulation to the loggia walls, adding external thermal insulation to the roof, changing solar reflectance of the finishing layer of the roof, changing windows, closing the loggias with operable glazing, adding solar shading in the loggias, and closing the open ground floor with operable glazing
Pazouki et al. [77]	2023	A seven-story academic building	Tehran, Iran	Multi-objective Decision-Making	DesignBuilder	Project profitability and maintenance costs	Energy-saving	-	The lighting system, roof insulation, wall external insulation, and PV installation
Abdeen et al. [20]	2024	Two-storey residential villa	UAE	Non-dominated sorting genetic algorithm (NSGA-II)	DesignBuilder	-	Annual energy consumption	-	Wall and roof insulation, glazing, infiltration rate, window shading, and setpoint and setback temperatures
Gea-Salim et al. [78]	2024	Heritage museum	Salta City, Argentina	multi-objective optimisation NSGA-II	EnergyPlus, JEPlus + EA software (OpenStudio SketchUp)	-	Annual energy consumption	Discomfort hours	Air renewals, internal wall insulation, external wall insulation, window type, and roof insulation
Dehghan and Amores [33]	2025	Residential building	Sari, Iran	NSGA-II	EnergyPlus, jEPlus + EA software	-	Primary energy consumption and greenhouse gas emissions	Indoor air quality, predicted percentage of dissatisfied, and visual discomfort hours	Heating and cooling setpoints, air infiltration rates, insulation types, window selections, airflow rates, and HVAC systems
Aruta et al. [35]	2025	Residential buildings	Italy	NSGA-II	MATLAB, DesignBuilder	Global cost saving and incentive variables	Primary energy saving	-	External wall thermal insulation, external roof thermal insulation, glazing, boilers, and air conditioning

Table 7. Selected previous studies of multi-objective optimisation retrofit (in chronological order).

Authors	Year	Case Studies	Location	Methods	Software	Decision Matrix Objectives			Energy Efficiency Measures (EEMS)
						Economic	Environmental	Social
Asadi et al. [53]	2012	Semi-detached house (one family) built in 1945	Portugal	Objective function; Thchebycheff programming formulation change; building data from the Portuguese building thermal code (RCCTE)	Bintprog function in MATLAB	Investment cost	Energy for heating, cooling, and water heating	-	Window type: external wall insulation, roof insulation, and solar collector type
Asadi [54]	2012	Residential building	Portugal	NSGA-II	TRNSYS, GenOpt and a Tchebycheff optimisation technique developed in MATLAB	Retrofit cost	Energy savings	Thermal comfort	External wall insulation materials, roof insulation materials, window type, and solar collector type.
Asadi et al. [25]	2014	School building	Coimbra, Portugal	Genetic Algorithm and Artificial Neural Network	TRNSYS, MATLAB	Retrofit cost	Energy consumption	Thermal discomfort hours	Building envelope parameters (including external wall and roof insulation materials, and window type) and the HVAC system type.
Shao et al. [55]	2014	Three-storey office building	Aachen, Germany	Non-dominated Sorting Genetic Algorithm-II (NSGA-II)	House of Quality	Initial investment cost (IC)	Annual operational energy, annual emission GWP, annual energy consumption, envelope air leakage, and climate	Indoor air quality; thermal comfort	Wall insulation, roof insulation, floor insulation, renewable insulation materials, improve building tightness, glazing insulation, advanced envelope technologies, heating/cooling system, building automation system, and photovoltaics
Penna et al. [26]	2015	12 sets of residential buildings	Italy	Genetic Algorithm-II (NSGA-II)	MATLAB TRNSYS 17	Investment cost (IC), net present value (NPV), energy cost (EC), maintenance cost (MC), replacement cost (RC), residual value (RV), and government incentives	Energy performance heating	Weighted discomfort time (WDT)	Wall insulation, roof insulation, floor insulation, glazing system, heating generator, and mechanical ventilation with heat recovery
Carli et al. [56]	2015	Public building	Bari, Italy	Pareto frontier	MATLAB with the Global Optimization Toolbox	the cost and its payoff	Reduction in electrical energy consumption, reduction in methane consumption, reduction in water consumption	Increase in occupants’ internal comfort	Energy efficiency, sustainability, and thermal comfort
Solmaz et al. [57]	2016	Public school building	Izmir, Turkey	Sensitivity analysis; Multi-objective optimisation	Sketch-up Open Studio; Simlab; MATLAB; GenOpt	Net present value (NPV)	Heating and cooling savings	-	Exterior walls, windows, shading, ground floor, and roof
Almeida [58]	2016	School buildings	Portugal	Multi-objective optimisation	DesignBuilder and ANN (Artificial Neural Networks)	Life-cycle cost (LCC)	Energy efficiency	Occupants’ thermal comfort	External wall U values, roof U values, window U values, the total solar energy transmittance of windows, and air change rate (ACR)
Camporeale et al. [59]	2017	Housing blocks	Seville	Genetic algorithm and Pareto rank	Rhinoceros, Grasshopper, and Galapagos; EnergyPlus through DIVA	Net present value (NPV)	Heating and cooling demand	-	Building shape, envelope U values, and window-to-wall ratio (WWR)
Kim et al. [60]	2017	General hospital building	Korea	Multi-objective optimisation	ILOG CPLEX STUDIO 6.0	Retrofit cost	Energy savings	-	Sets of external wall windows and insulation, roof insulation, and solar energy collectors
Ascione [61]	2017	Educational building	Benevento, South Italy	NSGA-II smart exhaustive sampling; Cost-optimal analysis	Energy Plus and MATLAB	Investment cost, global cost	Minimise thermal energy demand for space heating (TEDh) and space cooling (TEDc); annual percentage of discomfort hours (DH)	Thermal comfort	Thermal envelope, HVAC systems and equipment, and renewable energy sources
Jafari and Valentin [24]	2018	Ranch-style home	Albuquerque, NM	A genetic algorithm optimisation	eQuest (Quick Energy Simulation Tool)	Initial investment cost (IC), energy consumption cost (EC), maintenance and replacement cost (MR), property tax (TX), and resale value benefit (RV)	-	-
Bandera et al. [62]	2018	University building	Spain	NSGA-II	Energy Plus jEPlus	-	Minimum energy demand required for heating and cooling, minimise the exergy required, and maximise the energy available	-	Façade, roof, roof skylight, and windows
Bonamente et al. [63]	2018	Fire station	Cuneo, Italy	NSGA-II	Termolog EpiX8	Minimise discounted cost	Minimise thermal energy consumption, minimise electric energy consumption, and minimise GHG emissions	Maximise comfort level	Building insulation, heat generator, thermal distribution system, terminal units, heating control system, electrical systems, solar thermal collectors, and photovoltaic plant
Bosco et al. [64]	2018	Office building	Rome, Italy	NSGA-II	IDA-ICE and MOBO	Total investment cost	Annual total energy consumption	Annual discomfort hours	Four possible wall insulations, five roof insulations (in three possible thicknesses), and five possible windows, with different set-points for the heating curve
Rosso et al. [30]	2020	Residential building, three floors	Rome, Italy	Active-archive Non-dominated Sorting Genetic Algorithm-II (aNSGA-II)	EnergyPlus; Python version 3.6.8	Investment cost (IC) and energy cost (EC)	Energy demand and CO2 emissions	-	Glazing system, opaque vertical envelope insulation system, opaque horizontal envelope insulation system, opaque envelope finishing, layer optics characteristics, solar shading, sun space, closing the courtyard, closing the ground floor, solar thermal collectors, photovoltaic panels, and tilt angle of the thermal and photovoltaic panels
Pilechiha et al. [28]	2020	The office building’s windows	Tehran, Iran	Pareto Frontier and a weighted sum	Daylighting: Grasshopper plug-ins: Ladybug and Honeybee Energy simulation: Energy Plus and Open Studio	-	Minimise energy consumption and maximise the daylight	Visual comfort (absence of glare)	Wall materials, glazing systems, and window size
Rogeau et al. [65]	2020	Virtual building stocks	Rhone district in France	Single objective optimisation	R and RStudio version 3.5.3, and solved with the ILOG IBM CPLEX solver	Net present cost (NPC) and profitability	Energy demand and reduction in GHG emission	-	Building envelope and building heating systems
Aghamolaei et al. [66]	2020	Typical dwellings	Yazd, Iran	Parametric Sensitivity Analysis (PSA) multi-objective optimisation NSGA II	JEPlus and JEPlus-EA	-	Greenhouse gas emissions (GHG)	Improving indoor thermal comfort	-
Amani [67]	2020	Residential apartment building	Tehran, Iran	Multi-objective optimisation NSGA-II Euclidean distance method	DesignBuilder SimaPro	-	Minimise annual energy consumption and minimise global warming potential	-	Insulation system of the external wall
Amiri [68]	2020	A three-story residential building	Montreal, Canada	Bi-objective optimisation	Analytic hierarchy process (ANP) for multiple criteria decision-making (MCDM)	Minimise cost	Maximise utility	-	Passive energy technologies (e.g., window improvement, external wall insulation, internal wall insulation, vapour barrier, and weather barrier)
Ciardiello et al. [69]	2020	Residential apartment block	Rome, Italy	Active-archive non-dominated sorting genetic algorithm (aNSGA-II)	SketchUp, Open Studio, Energy Plus	Annual energy cost and investment cost	Total energy demand and CO2 emissions	-	Insulation on the opaque envelope and transmittance of the glazing system, the optic characteristics of the finishing layer of the external envelope, the presence or absence of brise-soleil and sunspaces on the loggias and in the greenhouse in the courtyard, and photovoltaic and solar thermal panels.
Chang et al. [70]	2020	Two apartments and two wooden houses	North Sumida, Tokyo, Japan	Multi-objective optimisation under uncertainties	Grasshopper in Rhinoceros 3D, Honeybee plugin, MATLAB	Payback period	CO2 emission and energy balance	Discomfort hours	Building envelope design
Qu et al. [29]	2021	Late nineteenth-century Victorian house	United Kingdom	NSGA-II	Energy Plus	Initial investment cost and payback period	Energy savings and annual energy consumption	-	Glazing system, airtightness, and thermal insulation
Hong et al. [71]	2021	Low-rise office building	Shanghai, China	NSGA-II, Linear Programming Techniques for Multidimensional Analysis of Preference (LINMAP)	Rhinoceros, Grasshopper plugins, namely Ladybug and honeybee, Octopus	-	Minimising energy demand	Maximising daylight availability	Thermochromic glazing
Cao et al. [72]	2021	Residential building	Northern Anhui, China	Multi-objective optimisation	DesignBuilder	Retrofit cost	Energy saving	Thermal comfort	External wall insulation layer and setting the roof insulation layer)
Tavakolan et al. [34]	2022	Single-family residence 240m2	Iran	NSGA-II	Building simulation: Energy Plus NSGA-II; MATLAB	Net present value (NPV), investment cost (IC), present value of electricity, natural gas, renewable energy, discounted payback period (DPP)	Primary energy consumption (PEC)	-	Wall Insulations, roof insulations, improving airtightness, window glazing, heating system, cooling system, and photovoltaic panels
Merlet et al. [27]	2022	Social housing stock; three apartment blocks are representative of the whole building’s stock	Paris, France	NSGA-II integration of temporality in optimisation, implementation of sequencing, and implementation of phasing	EnergyPlus in Design Builder DEAP Phyton	Cost	Heating demand and overheating	-	Window properties and wall properties
Swedberg [73]	2022	Multi-family residential projects	Humid Continental Climates (Dfb subtype)	Multi-objective design optimisation (MODO) NSGA-II	EnergyPlus, jEPlus, and jEPlus+EA	-	Minimise heating demand	Overheating-degree-hour (OHDH)	Exterior wall, slab, roof, airtightness, glazing (SHCG), and overhang depth
Aram [74]	2022	Educational building	Tehran, Iran	TOPSIS for retrofit measures	jEPlus, DesignBuilder	Minimise investment cost	Minimise heating and cooling load	-	Building envelopes, insulation, renewable energy, intelligent devices for managing energy consumption, indoor quality, and energy-efficient appliances
Abdelaziz et al. [75]	2023	Villa archetype	Egypt	Multi-objective optimisation	-	Life-cycle cost (LCC)	Life-cycle carbon footprint (LCCF)	-	Insulating external walls and roofs, upgrading existing windows, efficient lighting system, HVAC system, and solar PV cells
Ciardiello et al. [76]	2023	Social housing building	Rome, Italy	Active-archive non-dominated Sorting Genetic Algorithm (aNSGA-II)	Rhinoceros, Grasshopper, EnergyPlus	Investment and operational costs	Minimising energy consumption and CO2 emissions	-	Adding mid and internal thermal insulation to the external cavity walls, adding internal thermal insulation to the loggia walls, adding external thermal insulation to the roof, changing solar reflectance of the finishing layer of the roof, changing windows, closing the loggias with operable glazing, adding solar shading in the loggias, and closing the open ground floor with operable glazing
Pazouki et al. [77]	2023	A seven-story academic building	Tehran, Iran	Multi-objective Decision-Making	DesignBuilder	Project profitability and maintenance costs	Energy-saving	-	The lighting system, roof insulation, wall external insulation, and PV installation
Abdeen et al. [20]	2024	Two-storey residential villa	UAE	Non-dominated sorting genetic algorithm (NSGA-II)	DesignBuilder	-	Annual energy consumption	-	Wall and roof insulation, glazing, infiltration rate, window shading, and setpoint and setback temperatures
Gea-Salim et al. [78]	2024	Heritage museum	Salta City, Argentina	multi-objective optimisation NSGA-II	EnergyPlus, JEPlus + EA software (OpenStudio SketchUp)	-	Annual energy consumption	Discomfort hours	Air renewals, internal wall insulation, external wall insulation, window type, and roof insulation
Dehghan and Amores [33]	2025	Residential building	Sari, Iran	NSGA-II	EnergyPlus, jEPlus + EA software	-	Primary energy consumption and greenhouse gas emissions	Indoor air quality, predicted percentage of dissatisfied, and visual discomfort hours	Heating and cooling setpoints, air infiltration rates, insulation types, window selections, airflow rates, and HVAC systems
Aruta et al. [35]	2025	Residential buildings	Italy	NSGA-II	MATLAB, DesignBuilder	Global cost saving and incentive variables	Primary energy saving	-	External wall thermal insulation, external roof thermal insulation, glazing, boilers, and air conditioning