Forecasting the Estimate Completion Time of Syrian Building Projects Using Earned Value Management and Artificial Intelligence Dr. Samah MakKieh

Abstract

The time factor is one of the most important factors affecting on construction projects, and it needs effective techniques to manage it. The activities during the implementation phase take longer than the expected periods, Projects may fail and stop without being completed. The localization of the EVM methodology is a solution to this phenomenon; As it excels in the field of analyzing and evaluating the current time performance of the project, but its accuracy in forecasting the Estimate completion time of the project is insufficient, so this research aims to improve the performance of the EVM in forecasting the completion time of building projects using artificial neural networks.

The Education College and The Dentistry College at Tishreen University in Lattakia were selected as a case study. The network's inputs were determined based on: the values of performance indicators, the Project's Planned duration, the time variance, the Earned duration and the Earned schedule variables.

 The network was trained in several structural designs and the design corresponding to the lowest error was chosen as the best structure.

The training results indicated that the structure (8.6.1) represents the optimal network for forecasting the estimate completion time for the project. Finally, the optimum network was tested on fifteen samples that had not been previously trained. The test results proved the accuracy of the neural networks in forecasting.