Towards a systems engineering approach to tendering design in construction
Doctoral thesis, 2025
The tender phase is a time-constrained and business-critical stage in which the contractor uses its technical expertise to interpret and transform project requirements into a building concept. The growing volume and complexity of requirements create a need for a more systematised and automated approach to requirement analysis and conceptual design. One established approach to achieving more systematic requirement management, practised in other industries, is systems engineering. This research explores how data-driven and artificial intelligence-supported methods can facilitate such an approach in construction, focusing on requirements analysis, verification data, and the design process.
The research is structured into three parts, each addressing one research question. The first part investigates how project requirements can be automatically extracted, digitised, and analysed to support specialists in tendering. A prototype for requirements analysis was developed and evaluated through a workshop and two surveys. The second part explores production data for requirements and quality verification. It includes an assessment of data standardisation, a survey of the current benefits of digital reporting, and an exploration of future possibilities for using digital production data for knowledge generation and verification. The third part addresses tender-phase conceptual design, applying set-based design and a genetic algorithm to reduce production costs and carbon emissions in building superstructures. Two requirement-driven design prototypes were developed and tested on a reference building to visualise trade-offs.
The results show that a data-informed and AI-supported tendering process can facilitate requirements analysis, enable benchmarking across projects, and generate insights from verification data. AI support can also enable parallel evaluation of design alternatives with visualised trade-offs. To further improve a systems engineering approach in construction, incorporating continuous requirements analysis, standardised verification data, traceable requirement-verification links, and a broader analysis of design alternatives is recommended.
The research contributes to a more systematised and AI-supported approach to tendering by demonstrating how digital tools can help contractors analyse extensive and complex requirements, utilise verification data from previous projects, and generate design proposals that are both cost- and carbon-informed. This opens new opportunities to make better-informed decisions, improve early-phase design quality, and increase competitiveness through data-driven workflows.
The research is structured into three parts, each addressing one research question. The first part investigates how project requirements can be automatically extracted, digitised, and analysed to support specialists in tendering. A prototype for requirements analysis was developed and evaluated through a workshop and two surveys. The second part explores production data for requirements and quality verification. It includes an assessment of data standardisation, a survey of the current benefits of digital reporting, and an exploration of future possibilities for using digital production data for knowledge generation and verification. The third part addresses tender-phase conceptual design, applying set-based design and a genetic algorithm to reduce production costs and carbon emissions in building superstructures. Two requirement-driven design prototypes were developed and tested on a reference building to visualise trade-offs.
The results show that a data-informed and AI-supported tendering process can facilitate requirements analysis, enable benchmarking across projects, and generate insights from verification data. AI support can also enable parallel evaluation of design alternatives with visualised trade-offs. To further improve a systems engineering approach in construction, incorporating continuous requirements analysis, standardised verification data, traceable requirement-verification links, and a broader analysis of design alternatives is recommended.
The research contributes to a more systematised and AI-supported approach to tendering by demonstrating how digital tools can help contractors analyse extensive and complex requirements, utilise verification data from previous projects, and generate design proposals that are both cost- and carbon-informed. This opens new opportunities to make better-informed decisions, improve early-phase design quality, and increase competitiveness through data-driven workflows.
Tendering
Conceptual design
Requirement analysis
Systems engineering
Production data
Artificial Intelligence
Verifications
Author
Linda Cusumano
Chalmers, Architecture and Civil Engineering, Construction Management
Optimising cost and carbon emissions in the conceptual building design with genetic algorithms
Engineering Materials, Structures, Systems and Methods for a More Sustainable Future,;(2025)p. 1620-1625
Book chapter
Current benefits and future possibilities with digital field reporting
International Journal of Construction Management,;Vol. 25(2025)
Journal article
Clustering techniques and keyword extraction with large language models for knowledge discovery in building defects data
Construction Innovation,;Vol. 25(2025)p. 76-97
Journal article
Natural language processing as work support in project tendering
Current Perspectives and New Directions in Mechanics, Modelling and Design of Structural Systems,;(2022)p. 1583-1588
Paper in proceeding
Cusumano, L., Olsson, N, Granath, M, and Rempling, R. Guiding early building design toward lower carbon emissions through set-based design and genetic algorithm optimisation.
När ett byggföretag ska lämna ett anbud på ett nytt projekt har det ofta mycket kort tid på sig att ta fram en första stomförslag. Under den här fasen måste stora mängder beställarkrav tolkas och omsättas till en fungerande byggnad. Kraven kan vara många och komplexa, och dagens analysmetod är manuella, något som både tar tid och riskerar att viktiga detaljer förbises.
I min forskning undersöker jag hur metoder från systems engineering, tillsammans med artificiell intelligens och dataanalys, kan användas för att göra den här tidiga fasen mer effektiv och tillförlitlig. Avhandlingen omfattar tre delar, där den första fokuserar på beställerkrav ocg hur analysen av dessa kan automatiseras. Den andra delen fokuserar på produktionsdata i form av kvalitetsproblem och besiktningsanmärkningar, och hur den sortens information kan nyttjas för att generera kunskap till nya projekt. Den tredje och sista delen undersöker hur själva valet av stomme i anbudsskedet kan göras mer datainformerat, så att viktiga projektkrav - så som låga klimatutsläpp och rimliga produktionskostnader - kan uppfyllas.
Resultaten visar att datadrivna och AI-stöttade arbetssätt kan minska risken för misstag och ge projektteam bättre beslutsunderlag. Genom att tydligare se konsekvenserna av olika designval redan fi anbudsskedet kan entreprenörer både stärka sin konkurrenskraft och bidra till mer hållbara byggprojekt.
I min forskning undersöker jag hur metoder från systems engineering, tillsammans med artificiell intelligens och dataanalys, kan användas för att göra den här tidiga fasen mer effektiv och tillförlitlig. Avhandlingen omfattar tre delar, där den första fokuserar på beställerkrav ocg hur analysen av dessa kan automatiseras. Den andra delen fokuserar på produktionsdata i form av kvalitetsproblem och besiktningsanmärkningar, och hur den sortens information kan nyttjas för att generera kunskap till nya projekt. Den tredje och sista delen undersöker hur själva valet av stomme i anbudsskedet kan göras mer datainformerat, så att viktiga projektkrav - så som låga klimatutsläpp och rimliga produktionskostnader - kan uppfyllas.
Resultaten visar att datadrivna och AI-stöttade arbetssätt kan minska risken för misstag och ge projektteam bättre beslutsunderlag. Genom att tydligare se konsekvenserna av olika designval redan fi anbudsskedet kan entreprenörer både stärka sin konkurrenskraft och bidra till mer hållbara byggprojekt.
When a construction company prepares a tender for a new project, it often has very limited time to develop an initial structural concept. During this phase, large amounts of client requirements must be interpreted and translated into a feasible building design. These requirements can be numerous and complex, and today’s analysis methods are largely manual—something that takes time and increases the risk of overlooking important details.
In my research, I explore how methods from systems engineering, together with artificial intelligence and data analytics, can make this early phase more efficient and reliable. The thesis comprises three parts. The first focuses on client requirements and how their analysis can be automated. The second examines production data—such as quality issues and inspection remarks—and how this type of information can be used to generate knowledge for future projects. The third and final part investigates how the choice of structural system in the tendering phase can be made more data-informed, enabling essential project requirements—such as low carbon emissions and reasonable production costs—to be met.
The results show that data-driven and AI-supported approaches can reduce the risk of errors and provide project teams with better decision support. By making the consequences of different design choices more transparent during tendering, contractors can strengthen their competitiveness and contribute to more sustainable construction projects.
In my research, I explore how methods from systems engineering, together with artificial intelligence and data analytics, can make this early phase more efficient and reliable. The thesis comprises three parts. The first focuses on client requirements and how their analysis can be automated. The second examines production data—such as quality issues and inspection remarks—and how this type of information can be used to generate knowledge for future projects. The third and final part investigates how the choice of structural system in the tendering phase can be made more data-informed, enabling essential project requirements—such as low carbon emissions and reasonable production costs—to be met.
The results show that data-driven and AI-supported approaches can reduce the risk of errors and provide project teams with better decision support. By making the consequences of different design choices more transparent during tendering, contractors can strengthen their competitiveness and contribute to more sustainable construction projects.
Data-informed design with the help of artificial intelligence
Development Fund of the Swedish Construction Industry (SBUF) (13949), 2021-01-11 -- 2023-05-31.
NCC AB, 2021-01-11 -- 2023-05-31.
Driving Forces
Sustainable development
Innovation and entrepreneurship
Subject Categories (SSIF 2025)
Construction Management
Design
DOI
10.63959/chalmers.dt/5793
ISBN
978-91-8103-336-6
Doktorsavhandlingar vid Chalmers tekniska högskola. Ny serie: 5793
Publisher
Chalmers