The definition of the execution time of vertical residential buildings occurs, in
most cases, empirically and during the economic and financial feasibility study of the
project. This generates high risks and uncertainties regarding the deadline compliance.
However, there are tools of computational intelligence that can contribute to the
reduction of the degree of uncertainty, among them the fuzzy logic. This work aims to
develop a hybrid model to estimate, in a reasonable way, the construction period of
vertical residential buildings by associating factor analysis and fuzzy logic. For this it
was used a database from the Municipal Secretary of Urban Development of Belém
(SEURB) and the developers of the city, which are target of the case study, in the
period from 08/24/2015 to 09/11/2015. To build the model, in the first step, i.e., the
factor analysis, 127 projects were used; in the second step, the fuzzy inference, 71
projects were used; and, for the validation phase, 16 projects were used. Results
obtained showed a significant level of assertiveness, with accuracy of 75%, taking into
account an interval whose limits (upper and lower) were calculated in the mean
absolute percentage error (MAPE). The model delivered a prediction performance
superior to other models already accepted in literature. As theoretical contribution, the
work allows a better understanding of the uncertainties and risks regarding the
definition of the execution time of real estate projects, and also about how the modeling
can help entrepreneurs/construction campany and technicians to make better
decisions based on a hybrid model, considering the uncertain and complex
environment of the economic and financial feasibility study phase for future projects.
|
