An artificial neural network analysis of rework in sustainable buildings

Abstract

Abstract Sustainable buildings were introduced to lessen the negative impact of buildings on their users, occupants, communities, and the environment, and to increase contractors’ productivity and profit margins. However, despite the prospects and benefits of sustainable buildings, instances of reworking in buildings have been recorded by construction companies. This research developed a questionnaire instrument that included 20 determinants of rework in sustainable buildings, measured on a 5-point Likert scale. An artificial neural network model was used to predict the determinants of rework in sustainable buildings. The data revealed that sustainable buildings are delayed by more than 15% and cost overrun by more than 20%. Rework in sustainable buildings is about 15%. The correlation results revealed that rework will increase project costs by up to 29% and completion time by about 27%. The model revealed that lack of performance specification; lack of skilled workers with green materials and components; poor application of industrialised building systems; lack of project governance; absence of sustainable construction managers; poor welfare; and poor adoption of building information modelling are the main predictors of reworks. The research will provide references for all stakeholders involved in the procurement of sustainable buildings. This study is foremost to examine reworks in sustainable buildings, as its findings hold critical theoretical and practical implications for future research on the sustainable building market.