Advancing Concrete Mix Proportion through Hybrid Intelligence: A Multi-Objective Optimization Approach.

Area of Science:

• Civil Engineering and Materials Science
• Computational Intelligence and Optimization

Background:

• Concrete mixture design is critical for structural performance and cost-effectiveness.
• Traditional methods often lack precision and adaptability for complex optimization goals.
• Advancements in computational intelligence offer new avenues for optimizing material design.

Purpose of the Study:

• To develop a novel hybrid intelligent modeling approach for concrete mixture design.
• To predict concrete mechanical properties and optimize mix proportions using AI.
• To create a practical, AI-based system for efficient concrete mix design.

Main Methods:

• Hybrid intelligent modeling integrating artificial neural networks (ANN) and genetic algorithms (GA).
• GA-based optimization of ANN parameters for enhanced prediction accuracy (R² > 0.95).
• Multi-objective optimization using Scipy library and GA-ANN to balance compressive strength and cost.

Main Results:

• Achieved high prediction accuracy for concrete mechanical properties (R² > 0.95, low RMSE and MAE).
• Successfully optimized concrete mix proportions for multi-objective goals, balancing strength and cost.
• Developed a user-friendly AI-based system for practical concrete mix design.

Conclusions:

• The hybrid GA-ANN approach provides a novel and accurate method for concrete mixture design.
• The developed AI system offers a practical solution for optimizing concrete properties and costs.
• This research advances the efficiency and precision of concrete mix design in the construction industry.