Abstract
The rising growth and the number of construction projects required to be monitored day by day, to increase in the performance , hence the success for these projects . Because there are many factors that affect construction projects performance , the use of the smart technology enables us to obtain massive amounts of data around the clock, this facilitates predicting construction failures and management regulation. The monitoring of project performance across sectors has changed as a result of the convergence of artificial intelligence (AI) and the Internet of Things (IoT). This aim to build system that tracks, analyzes, and optimizes project performance using IoT and AI technologies.
In this study, a model was proposed to predict the construction projects failures and monitor its performance by using IoT and AI. The proposed model consists of four stages: data collection, preprocessing stage, the prediction stage and the performance optimization stage. Data set was collected through sensors
, questionnaire, camera and REVIT . Then, a machine and deep learning as machine learning using Decision tree , machine learning using Linear Regression and proposed hybrid linear regression with Gradient Boosting Regresso, machine learning using support vector classifier and proposed hybrid SVC with linear regression , machine learning using support vector classifier and proposed hybrid SVC with genetic algorithm and Deep learning using MLP and a proposed hybrid MLP with linear regression. A pre- processed data to predict the construction projects performance . The results of the model SVC-Linear are 99,9 and Gradient Boosting-Linear with the same accuracy with the same accuracy of decision tree, while the accuracy of multi- layer perception is 80 and hypered with linear regression is 67.8 and the worst accuracy of Gradient Boosting is 62.
It seen that most construction projects fail to meet the requirement of the performance and hence a suitable maintenance form and difference scenario regard each projects that failed to maintain the requirement , these scenario (Accept ,mitigate , prevent , predict) , in order to select the optimal selection two algorithm was used reptile search algorithm and particle swarm algorithm. both algorithm select the same scenario but particle swarm algorithm select the scenario with minimum effect .
