In the ever-evolving landscape of construction management, accurate cost estimation is a cornerstone for project success. A recent study spearheaded by Rui Wang from Universiti Malaya sheds light on a groundbreaking approach combining machine learning techniques with optimization algorithms. Published in the ‘Journal of Asian Architecture and Building Engineering’, this research introduces a hybrid model that promises to revolutionize how construction professionals estimate costs at the conceptual stage.
Wang’s research addresses a critical gap in the construction industry: the reliance on rough estimates that often lead to financial pitfalls. “Inaccurate initial estimates can strain project budgets and compromise financial viability,” Wang explains. This new model utilizes a hybrid Dung Beetle Optimizer (DBO) and Back-Propagation Neural Network (BPNN) to enhance the precision of cost estimations, particularly during the preliminary design phase.
The study meticulously identifies 20 key input variables that significantly influence cost estimation, employing techniques like SHAP analysis and correlation matrices. By analyzing a dataset comprising 117 general building projects through MATLAB simulations, the DBO+BPNN model emerged as the frontrunner, outperforming other combinations, including Genetic Algorithm (GA) and Particle Swarm Optimization (PSO) with BPNN.
The implications of this research extend beyond academic interest; they carry substantial commercial potential for the construction sector. Enhanced cost estimation accuracy means that firms can make more informed decisions, leading to better budget management and reduced financial risks. “This model not only provides reliable cost information but also enhances the likelihood of project success,” Wang asserts, highlighting its practical application in real-world scenarios.
As the construction industry faces increasing pressure to deliver projects on time and within budget, innovations like Wang’s model could become indispensable tools for project managers and stakeholders alike. By integrating advanced machine learning techniques into cost estimation processes, the industry can move toward a more data-driven future, minimizing reliance on subjective judgment and improving overall project outcomes.
This research not only illustrates the potential of machine learning in construction but also sets a precedent for future developments in the field, encouraging further exploration into optimizing project management processes. As the construction landscape continues to evolve, such advancements will be crucial for maintaining competitiveness and ensuring financial sustainability.
