International Research Journal of Engineering and Technology (IRJET)
e-ISSN: 2395-0056
Volume: 11 Issue: 09 | Sep 2024
p-ISSN: 2395-0072
www.irjet.net
Using an Artificial Neural Network to Predict Construction Materials Prices in Khartoum State Elkhider B. E. Mohamed1,2, Mohamed A. M. Elsawi 3, Mazin A. B. Mahmoud3, Lina K. M. Abdelmahmoud3 1Assistant Professor, Department of Civil Engineering, University of Science and Technology, Omdurman, Sudan 2Assistant Professor, Department of Civil Engineering, Elnasr University, Omdurman, Sudan
3 Student, Department of Civil Engineering, University of Science and Technology, Omdurman, Sudan ---------------------------------------------------------------------***---------------------------------------------------------------------
Abstract - As a result of the construction industry's large
Sudanese market pricing, one of the challenges encountered by this research is that cost assessment in Sudan is often done through internal cost analysis. A procurement issue involving thoroughly documented preceding work to research and evaluate the pre- and post-execution costs of construction projects is present.
expenditure, effective construction cost had become necessary. In order to estimate the expenses of building materials in Khartoum State, it is necessary to evaluate and determine whether ANNs is a viable instrument. The aim of this study is to evaluate the accuracy of artificial neural networks (ANNs) to calculate selected building materials prices in construction projects in Khartoum to decide ANNs ought to be used in the building sector. Four representative building projects across Khartoum State were selected, and the selected building materials expenses of those projects were then calculated. This information was gathered, loaded into MATLAB, and ANNs was created using the descriptive attributes as inputs and the corresponding building materials costs as the target output. The neural network then adjusted itself to try and predict the cost of the materials using the inputs that were provided. The cost created by the artificial neural network was compared to the actual cost after that. When the Neural Network outputs were compared to the real cost, it was discovered that there were absolute errors of 13.6%, 0, 2×10 -6 %, and 10-9 % for each project., the maximum absolute inaccuracy stated is 1.73%, and the mean absolute percentage error is 0.44%. The authors advocate using an Artificial Neural Network to estimate building costs and urge its development with a wider scope and larger sample size. Key Words: Khartoum.
Prices of materials may alter over the period of the construction project due to Sudan's irregular changes in hard currency exchange rates and the bank's lack of funds. The accuracy of the estimate by the project's completion date may be affected by these price changes. This study seeks to examine and determine whether an ANN is a useful tool for evaluating building project expenses in Khartoum State, Sudan. A further goal of the study is to decide whether artificial neural networks deserve to be used in the building sector.
2. LITERATURE REVIEW Numerous studies have been conducted on the issue of estimating construction costs using deep learning techniques. [3] Attempted to forecast the price of communication towers in Iraq using a mathematical model based on a multifactor linear regression technique. In this study, the author argued that the Multifactor Linear Regression Technique (MLR), a powerful mathematical tool for demonstrating the engineering interaction between dependent and independent variables, constituted the most effective forecasting model. The study concluded that (MLR) outperforms more traditional methods for predicting the costs of communication tower constructions, which are less specific and prone to ambiguity. 90.1% accuracy was achieved in terms of cost estimation [5].
ANN, Costs, Building Materials, MATLAB,
1. INTRODUCTION Due to of the construction industry's significant spending, effective building cost management and/or quantity surveying became essential issue. This occurs to ensure that a building project's resources are used to their maximum potential. Increasing cost performance is the fundamental objective of management for any building project, consequently the cost component typically takes priority [2].
According to [3], deep machine learning algorithms can provide high levels of accuracy and a decent level of confidence when used in prospective tower construction projects. Nevertheless, this project would accomplish its objectives through the usage of artificial neural networks (ANN). The study included data from a variety of construction companies [6]. The study hypothesized that the variables investigated would have a significant impact on the ANN's overall performance.
A construction project's performance has a high correlation with the capability to efficiently estimate spending, manage costs, and complete on time. [3][4] Mentioned that, to estimate material costs, a quantity surveyors must employ the greatest tools available. Due to the irregular nature of
© 2024, IRJET
|
Impact Factor value: 8.315
|
ISO 9001:2008 Certified Journal
|
Page 806