Predicting the construction duration of building projects using artificial neural networks

摘要

The accurate prediction of the duration of a construction project represents a critical factor for the feasibility study of this project. Employers are in an urgent need for reliable information about the construction duration in this early stage of the project. Such information can materially help project managers create a cash and material flow plan in a pre-set time. This paper aims to develop an artificial neural network (ANN) model for predicting the expected construction duration of building projects in its early stage, where no detailed planning is available. The MATLAB program was used as a suitable environment for developing the proposed model. The required field data was collected from 130 building projects in Egypt, which fall within the appropriate sample size. Testing the validity of the model clearly showed that it has a good prediction capability with a maximum error of 14%.