Developing a multi-output neural network model to predict construction duration of high-rise residential projects from the owners’ perspective | |
| Author | Hethusri, Vasantapu |
| Call Number | AIT Thesis no.CM-23-09 |
| Subject(s) | MATLAB Neural networks (Computer science) Construction projects |
| Note | A thesis submitted in partial fulfilment of the requirements for the degree of Master of Engineering in Construction, Engineering and Infrastructure Management |
| Publisher | Asian Institute of Technology |
| Abstract | This study is focused on identifying the important factors that affect the construction project duration from owners’ perspective and to predict the construction duration by developing an Artificial Neural Network model in MATLAB software. The study sets out to be held in Hyderabad, India. The data was collected from all the high-rise residential projects in and around Hyderabad. The study was undertaken by interviewing 2 experts from owners’ side regarding the real time construction duration estimation in Hyderabad. The factors that affect the duration were listed based on the literature review of previous researchers and case study interview. Then 5 experts were given to fill the expert validation form which has those factors, through which they are validated based on each expert. By that factors that affect the project duration from owners’ perspective as one of my objectives is satisfied. A neural network model has been developed in MATLAB, which will help the owner to predict the project duration by giving the inputs to the model. A configuration of 14-14-6 neural network was made, which has 14 input parameters and 6 output parameters. The model is trained with 14 hidden neurons with Sigmoid as its activation function. After training the model, it showed an accuracy of 90.44% which is higher than any of previous research models. A MAPE Value of 9.56% which is not a huge value. My model has 6 outputs in the model, predictions by each output have accuracies as follows, Excavation duration- 93.15%, Foundation duration-81.97%, Basement duration- 93.29%, Slab cycle duration-78.05%, Finishings duration-98.02%, Total project duration- 98.17%. Out of all the 6 outputs, total project duration is the output that is predicted at its close values. Finally, a predictor importance analysis was carried out in MATLAB. This analysis shows which of the 14 input parameters have more impact on the model. |
| Year | 2023 |
| Type | Thesis |
| School | School of Engineering and Technology |
| Department | Department of Civil and Infrastucture Engineering (DCIE) |
| Academic Program/FoS | Construction Engineering and Infrastructure Management (CM) |
| Chairperson(s) | Hadikusumo, Bonaventura H. W. |
| Examination Committee(s) | Tripathi, Nitin Kumar;Huynh, Trung Luong |
| Scholarship Donor(s) | AIT Scholarships |
| Degree | Thesis (M. Eng.) - Asian Institute of Technology, 2023 |