Politecnico di Milano Department of architecture, built environment and construction engineering (ABC) ABC-PhD Doctoral program KNOWLEDGE NETWORK FOR INNOVATION OF CONSTRUCTION SECTOR Increasing efficiency through process digitisation of the entire chain PhD candidate Claudio Mirarchi Supervisor: Prof. Alberto Pavan – Politecnico di Milano Tutor: Prof. Angelo Luigi Camillo Ciribini – Università degli studi di Brescia The Chair of the Doctoral Program Prof. Marco Scaioni – Politecnico di Milano XXXI Cycle March 2019 Politecnico di Milano Department of architecture, built environment and construction engineering (ABC) Via Giuseppe Ponzio 31, 20133 Milano (IT) ii To my parents that never limited my dreams and with their support have made this result possible. To Silvia that accompanied me in this long journey. iii iv Acknowledgements This work titled “knowledge network for innovation of construction sector” born from the collaboration between Assimpredil ANCE and Politecnico di Milano. It integrated both the industry and the academy of construction on a journey that allowed me to meet and work with countless people from both sides. A first and warm acknowledgement goes to my supervisor, Prof. Alberto Pavan, and to the director of Assimpredil ANCE, Gloria Domenighini, to have made possible the beginning and development of this research. Thanks to Alberto Pavan for his continuous support during these years and for the tireless willingness to share ideas and perspectives that have enormously contributed to shaping this work. Thanks to Gloria Domenighini for helping me to gain a different and privileged perspective toward the association and the industry and for paving the way in the work with the construction companies. Undoubtedly, the people who contributed to this work with suggestions and ideas are countless and it is impossible to thank them all individually. Nevertheless, some deserve a special acknowledgement. A heartfelt acknowledgement goes to Prof. Arto Kiviniemi who hosted and advised me during the research period at the University of Liverpool. Every single opportunity of discussion has allowed me to grow as a researcher and to improve my work. Thank also for accepting to review my thesis helping me to improve it up to its current form. An acknowledgement goes to all the research team of the University of Liverpool who made me feel at home during the months of research abroad. Special thanks go to Gulnaz Aksenova who tirelessly reviewed my work diving me useful suggestions to improve it. Thanks also to Sujesh Sujan for the continuous comparison that allowed me to see the work done from new angles. Thanks to all the Assimpredil ANCE staff who supported me in the many associative activities since the early days. Thanks to my tutors, Prof. Bruno Daniotti and Prof. Angelo Camillo Luigi Ciribini, for having reviewed and directed my work. Thanks to Prof. Žiga Turk for accepting to review my work and for pushing me to improve it up to its current form with new ideas and suggestions. Thanks also to all the construction companies that I met during these years of research for having listened and hosted me. Special thanks go to Giovanni Deleo who has always kept the doors of his company opened to the research and who has listened to me on countless topics. Thanks to Yang Zou, Jie Sze Fam and Prof. Alice Chang-Richards of the University of Auckland who assisted me in the development of machine learning applications for the organisation of text documents. A heartfelt thanks also to all the friends and colleagues who supported me during this research and who contributed to making these years special, sometimes helping me to raise my head from the screen and look elsewhere to see better the direction to take. To all of you, thank you. This work, and what I am, would not be the same without all of you. i Claudio Mirarchi – Knowledge network for innovation of construction sector. Increasing efficiency through process digitisation of the entire chain Ringraziamenti Questo lavoro intitolato “knowledge network for the innovation of construction sector” nasce dalla collaborazione tra Assimpredil ANCE e Politecnico di Milano, integrando sia l’industria che l’accademia delle costruzioni in un percorso che mi ha permesso di conoscere e confrontarmi con innumerevoli persone di entrambe le estrazioni. Un primo e caloroso ringraziamento va al mio supervisor, Prof. Alberto Pavan, ed al direttore generale di Assimpredil ANCE, Dott.ssa Gloria Domenighini, per aver reso possibile l’inizio e lo sviluppo di questa ricerca. Grazie ad Alberto Pavan per il supporto continuo durante questi anni e l’instancabile voglia di condividere idee e prospettive che hanno enormemente contributo a plasmare questo lavoro. Grazie a Gloria Domenighini per avermi aiutato a guadagnare una prospettiva diversa e privilegiata verso l’associazione e l’industria e per avermi aperto la strada verso il rapporto con le imprese. Indubbiamente le persone che hanno contribuito a questo lavoro con suggerimenti e idee sono innumerevoli ed è impossibile ringraziarle tutte singolarmente in queste pagine. Ciò nonostante, alcune meritano un ringraziamento speciale. Un sentito ringraziamento va al Prof. Arto Kiviniemi per avermi ospitato e consigliato durante il periodo di ricerca presso l’università di Liverpool. Ogni singola occasione di confronto mi ha permesso di crescere come ricercatore e di migliorare il mio lavoro. Grazie anche per aver accettato di revisionare la mia tesi aiutandomi a migliorarla fino alla sua forma attuale. Un ringraziamento a tutto il gruppo di ricerca dell’università di Liverpool che mi ha fatto sentire come a casa durante i mesi passati all’estero. Un ringraziamento speciale va a Gulnaz Aksenova che ha instancabilmente revisionato il mio lavoro regalandomi utili consigli per migliorarlo. Grazie anche a Sujesh Sujan per il confronto continuo che mi ha permesso di vedere il lavoro svolto da nuove angolazioni. Grazie a tutto lo staff di Assimpredil ANCE che mi hanno supportato fin dai primi giorni nelle molteplici attività associative. Grazie ai miei tutor, Prof. Bruno Daniotti and Prof. Angelo Camillo Luigi Ciribini, per aver revisionato e indirizzato il mio lavoro. Grazie al Prof. Žiga Turk per aver accettato di revisionare questa tesi e per avermi spinto a migliorarla fino alla sua forma attuale con nuove idee e suggerimenti. Un ringraziamento va anche tutte le imprese di costruzioni che ho incontrato durante questi anni di ricerca per avermi ascoltato e ospitato. Un ringraziamento speciale va a Giovanni Deleo che ha sempre tenuto aperte le porte della sua azienda alla ricerca e che mi ha ascoltato sui temi più vari. Grazie a Yang Zou, Jie Sze Fam ed alla Prof. Alice Chang-Richards, dell’università di Auckland, che mi hanno affiancato nello sviluppo delle applicazioni di machine learning per l’organizzazione dei documenti testuali. Un sentito ringraziamento anche a tutti gli amici e colleghi che mi hanno supportato durante questa ricerca e che hanno contributo a rendere questi anni speciali, aiutandomi talvolta ad alzare la testa dallo schermo e guardare altrove per vedere meglio la direzione da prendere. A tutti voi, grazie di cuore. Questo lavoro, e ciò che sono, non sarebbero gli stessi senza di voi. ii Abstract The construction industry has long been characterised as a project-based business that delivers one-of-a-kind products. No two buildings are alike, though they may appear similar from the outside. Hence, design and construction choices are made under uncertainty and the role of designers and constructors is consequently burdened with huge responsibilities about the success or failure of the project and its impact on the environment. Design and construction are processes that run in a cycle way through a trial-and-error procedure because it is not possible thought through the entire picture in advance. They can be considered as iterative procedures composed by a creative stage and a mechanical stage where the hypothesis of the former are tested with reference to project requirements. Moreover, the inherent variables of the construction site that can include machine/human failures, adverse weather conditions, and disadvantageous geological conditions brings the site management to face deviations between the as-planned and as-built construction performance, requiring real-time decisions during the construction development. In this context, decisions usually rely upon the knowledge of experts that act, however, under uncertainty and with access to a limited amount of information. Several studies in the field of psychology demonstrated the difficulties of human brain in evaluating samples, as people’s intuition appears to satisfy the law of small numbers. Hence, a possible precaution to the above-mentioned issues is the use of computation intended as the introduction of techniques and technologies able to support the deciders through data driven suggestions. However, the introduction of digital means in the construction sector demonstrated to be all but easy and linear. Thus, the peculiarities of the construction sector impose the development of social analysis to evaluate possible frameworks able to support the introduction of digital technologies and techniques devoted to improve knowledge management and decision making processes. On the other hand, the need to provide computational analysis in the construction sector impose technological challenges requiring the integration of disparate data generated along the construction chain in a highly fragmented