engleski

Early Stage Construction Cost Prediction in Function of Project Sustainability

Construction project costs often reach values higher than planned. Accuracy in project cost estimation is one of the most important criteria for project success, even for its sustainability. The main idea of this research is to examine the relationship between realized cost and contracted cost values for residential buildings. The aim of the research is to determine the mathematical relationship between realized and planned costs in the project implementation phase by using a few mathematical methods and some machine learning methods in comparison to linear regression. This would enable validation of methods themselves by comparing and evaluating the obtained relevant parameters. Comparison would be performed on two levels, based on its general characteristics, as well as on the results of their application on the basis of 24 building reconstructions and new buildings by comparing the mean absolute percentage error (MAPE) and the determination coefficient (R2 ) using Predictive Modelling Software DTREG (pronounced D-T-Reg). The relationship of realized and planned costs will be determined for the building as a whole and for certain types of construction works. That relationship would enable more realistic budget planning of similar future projects. Cost overrun factors will be analysed for particular types of construction works, as well as the probability of their occurrence, and what measures should be undertaken to prevent or reduce them in similar future projects. The phenomenon known in project planning as "optimism bias" will be analysed in the context of research focus of exceeding the construction cost

Building Construction, Contracted Cost, Realized Cost, Predictive Modelling Software, Machine Learning

nije evidentirano