Akademik Veri Yönetim Sistemi
Tezin Türü: Doktora
Tezin Yürütüldüğü Kurum: Yıldız Teknik Üniversitesi, İnşaat Fakültesi, İnşaat Mühendisliği Bölümü, Türkiye
Tezin Onay Tarihi: 2025
Tezin Dili: İngilizce
Öğrenci: Ahmed Abomahdy
Asıl Danışman (Eş Danışmanlı Tezler İçin): Aslı Pelin Gürgün
Eş Danışman: Cenk Budayan
Açık Arşiv Koleksiyonu: AVESİS Açık Erişim Koleksiyonu
Özet:
In the construction project management industry, determining the final project cost involves considering various influencing factors. Among the different methods, Estimate at Completion (EAC) stands out as a crucial approach for projecting the project's final cost. The key advantage of EAC lies in its ability to incorporate both the likelihood of project performance and associated risks. Moreover, EAC proves highly beneficial for project managers as it helps identify critical issues throughout the project's progression and enables the determination of appropriate solutions. This study introduces innovative coupled intelligence models, specifically integrating the Firefly Optimization Algorithm (FFA) and the Archimedes Optimization Algorithm (AOA) with Adaptive Neuro-Fuzzy Inference Systems (ANFIS) and Artificial Neural Network (ANN) for modeling project construction estimates at completion. AOA and FFA are utilized to identify significant attributes affecting the EAC which is dependent variable. This approach underscores the efficacy of hybrid models as a novel predictive framework within this domain. A traditional ANN is designed as a benchmark to evaluate the predictive accuracy of the hybrid models. These models are built using historical data from construction projects performed in Taiwan covering the period from 2000 to 2007. The study aims to determine EAC and assess the trend change in the forecast model monitor. The primary objective is to establish a reliable trend of EAC estimates, aiding the construction companies in enhancing project cost control effectiveness. The results highlight that the FFA-ANN model shows superior performance over traditional ANN, ANFIS models and hybridized AOA-ANN, AOA-ANFIS and FFA-ANFIS models.