Digital library of construction informatics and information technology in civil engineering and construction
Digital library
Paper: lc3-2017-260
| Paper title: | Predicting Highway Projects' Actual Duration Using Neural Networks |
| Authors: | Euthimios Glymis, Athanasios Kanelakis, Georgios Aretoulis and Theodoros Mastoras |
| Summary: | The current study predicts the actual duration of highway projects, based on the initial planned schedule. Highway projects suffer from delays and deviations, which in Greece are more often associated with law disputes, project financing, archaeological findings, environmental issues and private land acquisition procedures. In this research, data were obtained from 37 road projects in Greece and the purpose was to estimate the construction duration, using an artificial neural network. The Fast Artificial Neural Network (FANN) Tool program was used. FANN based on the available data, identifies the optimal training algorithm. The training algorithm and the activation function with the lowest mean square error (MSE) are selected. In order to achieve the best possible solution, numerous trials were made, applying different input data combinations, different variations in the architecture of the network, and different data values were used. This paper presents the three more reliable and effective networks produced from the current study. The results indicated that Artificial Neural Networks, employing the appropriate parameters do provide a relatively high accuracy in predicting actual construction time and more specifically appear as one of the most optimal methods for actual highway construction time prediction. |
| Type: | regular paper |
| Year of publication: | 2017 |
| Keywords: | Highway Projects, Actual Project Duration, Estimated Project Duration, Artificial Neural Networks |
| Series: | jc3:2017 |
| Download paper: | /pdfs/LC3_2017_paper_260.pdf |
| Citation: | Euthimios Glymis, Athanasios Kanelakis, Georgios Aretoulis and Theodoros Mastoras (2017). Predicting Highway Projects' Actual Duration Using Neural Networks. Lean and Computing in Construction Congress (LC3): Volume I Ð Proceedings of the Joint Conference on Computing in Construction (JC3), July 4-7, 2017, Heraklion, Greece, pp. 691-697, http://itc.scix.net/paper/lc3-2017-260 |
hosted by University of Ljubljana
includes:
W78
includes:
W78
inspired by SciX, ported by Robert Klinc [2019]