Measuring BIM implementation: A mathematical modeling and artificial neural network approach

Evaluating the level of Building Information Modeling (BIM) implementation in construction firms is critical yet challenging in the absence of a quantitative method. This study addresses this gap. The study begins with a literature review that identified 27 BIM implementation enablers, followed by interviews with three firms to score their performance on each enabler. A mathematical model was developed to score a firm's BIM implementation based on each enabler's score. For each firm, 1 million random scenarios are generated to simulate alternative ways by which a firm's enablers' score can be improved. Subsequently, in each simulated scenario, the firm's BIM implementation score is calculated. The simulation results are incorporated into a feature-pairing neural network that has been designed specifically to provide a customized best course of action for each firm's further BIM adoption. The first contribution of this research is providing a comprehensive analysis of the dynamics and interconnectedness of factors influencing BIM adoption in AEC firms, offering insights into effective BIM adoption. The second contribution is proposing a novel quantitative approach for measuring the current level of BIM implementation and providing data-driven advice for steering the BIM implementation process. This research offers a practical contribution by providing companies with a tool to compute their BIM implementation score, allowing comparisons and benchmarking against competitors.