Optimization of Road Construction Project Resource Allocations Using Artificial Neural Network

Abstract

As a vital economic driver, the construction industry contributes an average of 8-10% to national GDP of developed countries, while stimulating growth, creating employment, and enabling cross-sector economic linkages. Despite its significance, the sector faces persistent challenges including low productivity, inefficient resource allocation, frequent project delays, and cost overruns. This research proposed an Artificial Neural Network (ANN) approach to optimize resource allocation in construction projects, addressing three key objectives: analyzing critical factors affecting resource allocation, identifying enabling factors for optimal resource distribution, and developing an ANN-based optimization model. The study employed mixed methods research, combining semi structured interviews and questionnaire surveys to examine key determinants of resource allocation, including budget constraints, project scale, timelines, financial management, and resource availability. The research also identified crucial enabling factors such as detailed planning, effective project management, and robust monitoring systems. Quantitative data analysis utilized descriptive statistics, Relative Importance Index (RII), and correlation methods. The developed ANN model, implemented in MATLAB R2019a, features a three-layer architecture (input, hidden, and output layers) specifically designed to predict equipment utilization in asphalt construction projects. The model demonstrated remarkable predictive accuracy, achieving a near-perfect regression value (R = 0.9999) and an extremely low Mean Squared Error (MSE = 9.26×10⁻¹⁶), with minimal deviations between predicted and actual values. These results confirm the model's capability to transform historical project data into reliable decision-support tools, offering significant improvements in resource planning accuracy and cost estimation efficiency. The findings present a valuable framework for construction project managers seeking to enhance resource allocation through advanced predictive analytics, ultimately contributing to improved project outcomes and sector-wide productivity gains.