Scholarlink

<p>This study represents an advanced predictive and optimization methodology for High-Performance Concrete (HPC) using quantum kernel support vector machines (QKSVM) and multi-objective optimization strategies that integrate nano-alumina and zeolites. This approach achieves high levels of predictive accuracy by exploiting quantum kernels for non-linear interactions in high dimensions, thereby surpassing conventional machine learning applications. Integrating SHapley Additive exPlanations (SHAP) complements Gradient Boosted Decision Trees to analytically inform interpretability in quantifying feature contributions. Bayesian Optimization with Q-Learning takes this mix design optimization advancements into higher levels by saving about 30% of experimental trials while attaining strength and durability targets. The multi-objective optimization framework that hybridizes Differential Evolution (DE) with Non-Dominated Sorting Genetic Algorithm-I (NSGA-I) strikes an optimum balance of compressive strength, durability, and environmental Impact, tapping up to 20% reductions in CO₂ outputs. Comparison experiment shows that the prediction accuracy and mix design efficiency, from which this model derives, improved by about 15 to 20%. The research in this area presents a data-driven, computationally efficient framework for the optimization of HPC formulations within a contribution to sustainable constructions. This is used building towards a scalable framework for quantum machine learning implementations in materials engineering toward more resilient and environmentally sustainable solutions in development infrastructures in process.</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

Predictive performance of nano-alumina and zeolite-based high-performance nano-engineered concrete: integrative application of quantum computing and machine learning with optimization techniques

  • 期刊论文

  • 发表日期：

    2025年10月3日

查看全文
  • Tejaswini G. Panse,
  • Monica N. Kalbande,
  • Rupali S. Balpande,
  • Yashika A. Gaidhani,
  • Yoginee S. Pethe,
  • Nilesh Shelke,
  • Vikrant S. Vairagade

摘要

This study represents an advanced predictive and optimization methodology for High-Performance Concrete (HPC) using quantum kernel support vector machines (QKSVM) and multi-objective optimization strategies that integrate nano-alumina and zeolites. This approach achieves high levels of predictive accuracy by exploiting quantum kernels for non-linear interactions in high dimensions, thereby surpassing conventional machine learning applications. Integrating SHapley Additive exPlanations (SHAP) complements Gradient Boosted Decision Trees to analytically inform interpretability in quantifying feature contributions. Bayesian Optimization with Q-Learning takes this mix design optimization advancements into higher levels by saving about 30% of experimental trials while attaining strength and durability targets. The multi-objective optimization framework that hybridizes Differential Evolution (DE) with Non-Dominated Sorting Genetic Algorithm-I (NSGA-I) strikes an optimum balance of compressive strength, durability, and environmental Impact, tapping up to 20% reductions in CO₂ outputs. Comparison experiment shows that the prediction accuracy and mix design efficiency, from which this model derives, improved by about 15 to 20%. The research in this area presents a data-driven, computationally efficient framework for the optimization of HPC formulations within a contribution to sustainable constructions. This is used building towards a scalable framework for quantum machine learning implementations in materials engineering toward more resilient and environmentally sustainable solutions in development infrastructures in process.