Abstract

Cost estimation is one of the vital processes in construction management that needs to be done early in any project in order to determine the project's budget. The accuracy of the cost estimate is a key factor in the success of construction projects since it enables project managers to successfully control the project’s expenses. Construction costs mainly consist of direct cost and indirect cost. Generally, indirect costs can be categorized into two types: site overheads and general overheads. In a construction project, overheads, particularly site overhead costs, make up a considerable portion of a contractor's budget. Accordingly, accurately estimating the site overheads of construction projects is a crucial task that needs to be done in order to manage projects efficiently. Therefore, the main objective of this research is to enhance the contractor’s ability to accurately predict the site overhead costs of construction projects in Egypt through identifying and analyzing the key factors influencing site overheads in the Egyptian construction industry. This study proposes three-stage ANN approach for predicting site overheads of construction projects. The first ANN model estimates the total site overhead percentage while the second and third ANN models then utilize both the predicted total site overheads percentage and existing project data to forecast the breakdown of site overhead across its different subcategories and across the project’s different construction phases, while incorporating both economic and non-economic variables. In order to form the model’s database, the major factors affecting the site overheads were first identified through an extensive literature review. These factors were project type, project location, project duration, contract type, project direct cost, client type, class of contracting company and lastly macroeconomic indicators such as inflation rate, interest rate and currency exchange rates. In addition, cost data from 55 real-life projects executed during the past 10 years were obtained to be used as a database for the learning process of the ANN model. Cost data for 5 new projects were then used to test each model. Model 1 had 2.75% MAE for training set and 3.9% for testing set. Model 2 had 2.62% MAE for training and 2.83% for testing while Model 3 had 2.12% MAE for training and 2.31% MAE for testing data set. Overall, the models performed well and can be considered a useful tool for the predicting the percentage of site overheads as well as the percentage of site overheads allocated to each subcategory and each construction phase. Thus, these models offer a valuable tool for contractors to enhance cost estimation, improve decision-making and mitigate financial risks.

School

School of Sciences and Engineering

Department

Construction Engineering Department

Degree Name

MS in Construction Engineering

Graduation Date

Winter 2-19-2025

Submission Date

9-11-2024

First Advisor

Dr. Ossama Hosny

Second Advisor

Dr. Elkhayam Dorra

Committee Member 1

Dr. Mohamed Mahdy

Committee Member 2

Dr. Ibrahim Abotaleb

Extent

130 P.

Document Type

Master's Thesis

Institutional Review Board (IRB) Approval

Not necessary for this item

Download

Share

COinS