نوع مقاله : مقاله پژوهشی
نویسندگان
1 گروه مهندسی عمران، دانشکده فنی و مهندسی، دانشگاه آزادا اسلامی واحد یزد، یزد، ایران
2 دانشجوی دکتری رشته عمران گرایش مهندسی و مدیریت ساخت، واحد یزد، دانشگاه آزاد اسلامی، یزد،
3 عضو هیئت علمی گروه عمران، دانشگاه آزاد اسلام یزد، یزد، ایران
چکیده
کلیدواژهها
موضوعات
عنوان مقاله [English]
نویسندگان [English]
Because the goal of project oversight is to make accurate decisions that can have a significant impact on project success, predicting project features becomes even more important. Experts believe that the vast majority of construction projects suffer from delays. Therefore, one of the most important features of construction projects is issues related to time. This article proposes a model to solve project scheduling problems to some extent. To this end, this study presents new applications of short-term long-term memory prediction (LSTM) models, which are a recursive neural network architecture. On the other hand, in order to compare and validate the LSTM method, the gateway learning model (GRU) is examined. Subsequently, the prediction results of the proposed models are compared and verified with the data of a real project. In this study, to predict the real data of the southern development project of Tehran Metro Line 6, which was completed in 1997, was used and to measure the accuracy, the root mean square measure was used. The results show that short-term long-term memory and proposed applications of the model can accurately predict project progress. Also, in the GRU forecast compared to LSTM, the square root of the mean squared decreased by about 30%.
کلیدواژهها [English]