Scholarlink

<p>This paper provides an amalgamated approach incorporating Building Information Modeling (BIM) with Artificial Intelligence (AI) for better cost and scheduling predictions in energy-efficient infrastructure projects. A total of 245 cases of actual projects involving transport infrastructure, residential infrastructure, and mixed infrastructure were used within the study, where machine learning algorithms such as LightGBM, XGBoost, and Random Forest were fine-tuned with six different metaheuristic optimization algorithms including Particle Swarm Optimization (PSO) and Genetic Algorithm (GA). The proposed approach demonstrated excellent predictive accuracy, with a LightGBM-PSO model achieving a schedule accuracy of 91.5% and a cost F1 value of 0.91. Results showed that incorporating factors related to project sustainability, including CO₂ emission rates and LEED rating, improved predictive accuracy, thus advocating the use of these factors within project risks associated with infrastructure projects. To acquire explainable outcomes, both global and instance-level attribute explanations were performed within the study with the use of a SHAP-value analysis tool.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Integrating Building Information (BIM) and artificial intelligence to enhance cost and schedule planning in energy-conscious infrastructure projects

  • 期刊论文

  • 发表日期：

    2025年12月18日

查看全文
  • Ahmad Habboush,
  • Bassam Elzaghmouri,
  • Ola Adnan Altiti

摘要

This paper provides an amalgamated approach incorporating Building Information Modeling (BIM) with Artificial Intelligence (AI) for better cost and scheduling predictions in energy-efficient infrastructure projects. A total of 245 cases of actual projects involving transport infrastructure, residential infrastructure, and mixed infrastructure were used within the study, where machine learning algorithms such as LightGBM, XGBoost, and Random Forest were fine-tuned with six different metaheuristic optimization algorithms including Particle Swarm Optimization (PSO) and Genetic Algorithm (GA). The proposed approach demonstrated excellent predictive accuracy, with a LightGBM-PSO model achieving a schedule accuracy of 91.5% and a cost F1 value of 0.91. Results showed that incorporating factors related to project sustainability, including CO₂ emission rates and LEED rating, improved predictive accuracy, thus advocating the use of these factors within project risks associated with infrastructure projects. To acquire explainable outcomes, both global and instance-level attribute explanations were performed within the study with the use of a SHAP-value analysis tool.