The RISCONA system: constructability appraisal through the identification and assessment of technical project risks sources
Paper i proceeding, 2019
In construction management, constructability and risk analysis have never been methodologically and computationally integrated, leading to non-optimal construction knowledge implementation, stakeholders’ cooperation, choice of construction method, and risk-driven perception of key managerial concepts. In this paper, a methodology unifying constructability and risk analysis is delineated, where: (1) risk sources are derived with unsupervised machine learning, (2) actual projects’ data are collected and suitably correlated with the derived risk sources, and (3) the appraisal of constructability through the data-correlated risk sources is modelled with supervised machine learning. As the culmination of this modelling, the prototype software application RISCONA (RIsk Source-based CONstructability Appraisal) is presented, as a tool that can help construction managers in their decision-making regarding constructability and risk analysis.
nonnegative matrix factorization
support vector machines
prototype software application
unsupervised and supervised machine learning