liu.seSearch for publications in DiVA
CiteExportLink to record
Permanent link

Direct link
Cite
Citation style
  • apa
  • ieee
  • modern-language-association-8th-edition
  • vancouver
  • oxford
  • Other style
More styles
Language
  • de-DE
  • en-GB
  • en-US
  • fi-FI
  • nn-NO
  • nn-NB
  • sv-SE
  • Other locale
More languages
Output format
  • html
  • text
  • asciidoc
  • rtf
The Development of Complex Intelligent Systems: A Study of Engineering and Project Management Approaches
Linköping University, Department of Management and Engineering, Project Innovations and Entrepreneurship. Linköping University, Faculty of Science & Engineering.ORCID iD: 0009-0008-1755-0988
2026 (English)Doctoral thesis, comprehensive summary (Other academic)
Abstract [en]

Advances in software and artificial intelligence (AI) are rapidly transforming society. Many of the systems that underpin everyday life, such as aircraft, vehicles, energy systems, and transportation infrastructure, are complex technical systems that are becoming increasingly software-intensive. As these systems gain the ability to make decisions, learn from data, and adapt to their environment, they evolve into what this thesis refers to as Complex Intelligent Systems (CoIS).

This thesis examines how the development of such systems is changing as software and AI grow in importance, with a particular focus on engineering and project management approaches. Traditionally, complex systems have been developed using systems engineering, often supported by model-based approaches to ensure structured, traceable, and predictable development. At the same time, development projects have typically been organized through phase-based Stage-Gate processes.

However, the growing integration of software and AI has led to the emergence of data-driven engineering approaches and more iterative and flexible agile practices. This thesis investigates how these different approaches interact in the development of CoIS.

The study is based on interviews with experts from the aviation, automotive, and naval industries, complemented by an in-depth case study in the automotive domain. The findings show that the development of CoIS rarely involves replacing established approaches with new ones. Instead, organizations need to combine multiple approaches simultaneously. Model-based approaches remain essential for safety-critical systems, while data-driven and agile approaches are required to support learning and adaptability. This creates fundamental tensions between predictability and control on the one hand, and learning and flexibility on the other. The thesis contributes to understanding these tensions by showing that it is not only merely a practical or methodological challenge, but rooted in fundamentally different underlying logics of development embedded in the approaches used.

The results also highlight the role of system architecture as a key mechanism for managing this complexity. By structuring systems appropriately, organizations can combine different development approaches without compromising safety or functionality. Overall, the main challenge is not choosing a single approach, but managing the interplay between multiple approaches across both the system and the organization.

This thesis contributes by providing an integrated perspective on the development of CoIS and by identifying the organizational and managerial challenges that arise when different development logics intersect. As CoIS development is still evolving, further research is needed to understand how organizations can develop the capabilities required to manage these dynamics over time and across industries.

Abstract [sv]

Framsteg inom mjukvara och artificiell intelligens förändrar snabbt vårt samhälle. Många av de system vi är beroende av i vardagen, till exempel flygplan, fordon, energisystem, och transportsystem, är komplexa tekniska system som utgör en viktig del av samhällets infrastruktur. Dessa system blir allt mer mjukvaruintensiva och börjar i allt högre grad kunna fatta beslut, lära sig från data och anpassa sig till sin omgivning. De utvecklas därmed till vad som i denna avhandling kallas komplexa intelligenta system (Complex Intelligent Systems, CoIS).

Den här avhandlingen undersöker hur utvecklingen av sådana system förändras i takt med att mjukvara och AI blir allt viktigare. Särskilt fokus ligger på hur arbetssätt inom ingenjörsutveckling och projektledning påverkas. Traditionellt har utvecklingen av komplexa system byggt på så kallad systems engineering, där modellbaserade metoder används för att planera, analysera och verifiera system på ett strukturerat och förutsägbart sätt. Samtidigt har utvecklingsprojekten ofta organiserats genom fasindelade projektmodeller med tydliga beslutspunkter, så kallade Stage-Gate-processer.

När mjukvara och AI integreras i allt fler system växer dock andra arbetssätt fram. Datadrivna metoder, där system förbättras genom analys av stora mängder data, blir allt vanligare. Samtidigt sprids agila arbetssätt inom projektledning, där utvecklingen sker mer iterativt och flexibelt. Avhandlingen undersöker hur dessa olika arbetssätt samspelar när komplexa intelligenta system utvecklas.

Studien bygger på intervjuer med experter inom flyg-, fordons- och marinindustrin samt en djupgående fallstudie inom fordonsindustrin. Genom dessa studier analyseras hur organisationer i praktiken kombinerar olika utvecklings- och projektledningsmetoder när nya tekniskt avancerade system tas fram.

Resultaten visar att utvecklingen av komplexa intelligenta system sällan innebär att gamla arbetssätt ersätts helt av nya. I stället behöver organisationer använda flera olika modeller samtidigt. Modellbaserade arbetssätt är fortfarande avgörande för säkerhetskritiska system, eftersom de gör det möjligt att säkerställa spårbarhet, transparens och förutsägbarhet i utvecklingen. Samtidigt kräver mjukvara och AI datadrivna och mer flexibla arbetssätt. Detta skapar en spänning mellan behovet av kontroll och förutsägbarhet å ena sidan, och behovet av lärande och anpassning å andra sidan.

Avhandlingen visar också att systemarkitekturen – hur ett system delas upp i olika delar och hur dessa samverkar – spelar en viktig roll för att hantera denna komplexitet. Genom att strukturera systemet på ett genomtänkt sätt kan organisationer kombinera olika utvecklingsmetoder utan att kompromissa med säkerhet eller funktion.

Ett centralt resultat är därför att den största utmaningen inte är att välja en metod för utveckling, utan att hantera samspelet mellan flera olika arbetssätt samtidigt. Organisationer behöver kunna balansera och samordna modellbaserade, datadrivna, agila och fasstyrda metoder både i själva systemet och i organisationen.

Avhandlingen bidrar till forskningen genom att ge en samlad bild av utvecklingen av komplexa intelligenta system. Den visar vilka organisatoriska och ledningsmässiga utmaningar som uppstår när olika utvecklingslogiker möts. Eftersom utvecklingen av sådana system fortfarande befinner sig i ett tidigt skede behövs mer forskning för att förstå hur organisationer kan utveckla arbetssätt och förmågor för att hantera dessa nya förutsättningar över tid och i olika branscher.

Place, publisher, year, edition, pages
Linköping: Linköping University Electronic Press, 2026. , p. 170
Series
Linköping Studies in Science and Technology. Dissertations, ISSN 0345-7524 ; 2530
Keywords [en]
Complex Intelligent Systems, Model-based, Data-driven, Agile, Stage-Gate, Artificial Intelligence
National Category
Software Engineering Artificial Intelligence
Identifiers
URN: urn:nbn:se:liu:diva-223371DOI: 10.3384/9789181186031ISBN: 9789181186024 (print)ISBN: 9789181186031 (electronic)OAI: oai:DiVA.org:liu-223371DiVA, id: diva2:2056047
Public defence
2026-05-29, A2, A-building, Campus Valla, Linköping, 09:15 (English)
Opponent
Supervisors
Available from: 2026-04-28 Created: 2026-04-28 Last updated: 2026-05-05Bibliographically approved
List of papers
1. Understanding the development of emerging complex intelligent systems
Open this publication in new window or tab >>Understanding the development of emerging complex intelligent systems
2024 (English)In: Journal of engineering and technology management, ISSN 0923-4748, E-ISSN 1879-1719, Vol. 72, article id 101815Article in journal (Refereed) Published
Abstract [en]

This paper explores the intricate emergence of complex and increasingly intelligent systems (CoIS) in the wake of new possibilities created by integrating artificial intelligence (AI) solutions, building on an analysis of the emergence of CoIS using perspectives of development and change. The findings, based on rich qualitative data collected through key informant interviews with reflective practitioners from aviation, automotive and naval system domains, indicate that firms facing the emergence of CoIS, need to build capabilities allowing several logics to co-exist in a newly evolving hybrid CoIS management logic.

Place, publisher, year, edition, pages
ELSEVIER, 2024
Keywords
Artificial Intelligence (AI); Complex intelligent systems (CoIS); Model -based methods; Data -driven methods; Agile
National Category
Business Administration
Identifiers
urn:nbn:se:liu:diva-203775 (URN)10.1016/j.jengtecman.2024.101815 (DOI)001221320500001 ()
Note

Funding Agencies|Wallenberg AI, Autonomous Systems and Software Program - Humanity and Society (WASP-HS) - Marianne and Marcus Wallenberg Foundation

Available from: 2024-05-31 Created: 2024-05-31 Last updated: 2026-04-28
2. Expanding data usage in systems: an empirical study of combined model-based and data-driven development in complex intelligent systems (CoIS)
Open this publication in new window or tab >>Expanding data usage in systems: an empirical study of combined model-based and data-driven development in complex intelligent systems (CoIS)
2025 (English)In: Proceedings of the Design Society 2025, Dallas, Texas, USA, 11–14 august 2025 / [ed] Gaetano Cascini, Cambridge University Press , 2025, Vol. 5, p. 1615-1624Conference paper, Published paper (Refereed)
Abstract [en]

AI is becoming an important part of complex products and systems (CoPS), transforming them into complex intelligent systems (CoIS), on which our society depends. Traditionally, system development relied on model-based approaches, and the emerging data-driven approaches offer new possibilities. This paper explores the intertwining of model-based and data-driven approaches in emerging CoIS through a comparative case study of their role in cloud-based automotive systems, which are part of the transportation system. The findings show that data-driven approaches not only complement model-based approaches but also play a pivotal role in the evolution of CoIS.

Place, publisher, year, edition, pages
Cambridge University Press, 2025
Series
Proceedings of the Design Society, ISSN 2732-527X
Keywords
artificial intelligence, complex intelligent systems, data driven, large-scale engineering systems, systems engineering (SE)
National Category
Artificial Intelligence Software Engineering
Identifiers
urn:nbn:se:liu:diva-223370 (URN)10.1017/pds.2025.10175 (DOI)2-s2.0-105022823665 (Scopus ID)
Conference
25th International Conference on Engineering Design (ICED25)
Note

Funding: This work was partially supported by the Wallenberg AI, Autonomous Systems and Software Program – Humanity and Society (WASP-HS) funded by the Marianne and Marcus Wallenberg Foundation.

Available from: 2026-04-28 Created: 2026-04-28 Last updated: 2026-04-28Bibliographically approved

Open Access in DiVA

fulltext(5391 kB)230 downloads
File information
File name FULLTEXT02.pdfFile size 5391 kBChecksum SHA-512
5ed88813448b006af443ff977a66770d5205a8e19077e29d59779008a3f9bff2db9c5863b5717aa2c3444c8155f6ca84622c9f5f63b3d2e91b8826042151a270
Type fulltextMimetype application/pdf
Order online >>

Other links

Publisher's full text

Authority records

Balachandran, Appu

Search in DiVA

By author/editor
Balachandran, Appu
By organisation
Project Innovations and EntrepreneurshipFaculty of Science & Engineering
Software EngineeringArtificial Intelligence

Search outside of DiVA

GoogleGoogle Scholar
Total: 230 downloads
The number of downloads is the sum of all downloads of full texts. It may include eg previous versions that are now no longer available

doi
isbn
urn-nbn

Altmetric score

doi
isbn
urn-nbn
Total: 932 hits
CiteExportLink to record
Permanent link

Direct link
Cite
Citation style
  • apa
  • ieee
  • modern-language-association-8th-edition
  • vancouver
  • oxford
  • Other style
More styles
Language
  • de-DE
  • en-GB
  • en-US
  • fi-FI
  • nn-NO
  • nn-NB
  • sv-SE
  • Other locale
More languages
Output format
  • html
  • text
  • asciidoc
  • rtf