Scholarlink

<p>Compressive strength is one of the key factors defining high-performance concrete (HPC), with high bearing loads being supported with reduced use of materials. With high-performance concrete utilizing complex materials such as fly ash and slag, HPC is characterized by increased durability, early age, and high crack resistance. In forecasting HPC compressive strength, Adaptive Boosting Regression (ADAR) and Decision Tree Regression (DTR), along with complex algorithms such as the Improved Manta-Ray Foraging Optimizer (IMRF) and Dingo Optimization Algorithm (DOA), have been utilized in this work. Unlike previous studies that primarily applied single machine learning models or conventional optimization methods, this study introduces novel hybrid models (ADIM, ADDO, DTIM, and DTDO) that integrate ADAR and DTR with advanced bio-inspired optimizers. This combination allows for superior handling of the nonlinear relationships in HPC mix designs while incorporating sustainability-related variables such as embodied CO₂ and energy use. Out of all the analyzed models and approaches, the best performance was achieved by the ADIM model, with an R² value of 0.994 and a low RMSE of 11.321, outperforming baseline ADA and DTR models. These results highlight the methodological advancement of integrating boosting algorithms with metaheuristic optimization, offering both improved predictive accuracy and practical applicability in optimizing HPC mix designs for stronger and more sustainable concrete.</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

Investigating the impact of optimization methods in enhancing the automated prediction of compressive strength of high-performance concrete

  • 期刊论文

  • 发表日期：

    2025年10月3日

查看全文
  • Fanglu Pan

摘要

Compressive strength is one of the key factors defining high-performance concrete (HPC), with high bearing loads being supported with reduced use of materials. With high-performance concrete utilizing complex materials such as fly ash and slag, HPC is characterized by increased durability, early age, and high crack resistance. In forecasting HPC compressive strength, Adaptive Boosting Regression (ADAR) and Decision Tree Regression (DTR), along with complex algorithms such as the Improved Manta-Ray Foraging Optimizer (IMRF) and Dingo Optimization Algorithm (DOA), have been utilized in this work. Unlike previous studies that primarily applied single machine learning models or conventional optimization methods, this study introduces novel hybrid models (ADIM, ADDO, DTIM, and DTDO) that integrate ADAR and DTR with advanced bio-inspired optimizers. This combination allows for superior handling of the nonlinear relationships in HPC mix designs while incorporating sustainability-related variables such as embodied CO₂ and energy use. Out of all the analyzed models and approaches, the best performance was achieved by the ADIM model, with an R² value of 0.994 and a low RMSE of 11.321, outperforming baseline ADA and DTR models. These results highlight the methodological advancement of integrating boosting algorithms with metaheuristic optimization, offering both improved predictive accuracy and practical applicability in optimizing HPC mix designs for stronger and more sustainable concrete.