This repository contains all materials related to my Master's Thesis titled "Generative Artificial Intelligence and Optimisation Framework for Sustainable Concrete Mixture Design", including the thesis report, presentations, research papers, dataset, and associated code files used for experiments and analysis.
-
Thesis Report
234156019_Teiboklang_Chyne_MTP_THESIS_Report.pdf
Final report of the Master's Thesis submitted to the department. -
Thesis Presentation
234156019_Teiboklang_Chyne_THESIS_Presentation.pptx
Slide deck presented as part of the thesis defense. -
Paper Submitted to RILEM Youth Symposium 2025
Rilem-Youth-Symposium-2025-Paper.docx
-
Paper Submitted to I-4AM’26
iAM26-Paper.pdf
- Original Dataset Used
Wiley.csv
The primary dataset used for training and evaluation in the thesis.
-
Firefly Optimization Implementation
firefly_optimization.ipynb
Optimizes CatBoost hyperparameters using the Firefly Algorithm. -
Distance Metric Comparison for Firefly
firefly_distance_metric_optimization.ipynb
Analyzes the impact of various distance metrics on Firefly optimization. -
Analysis of Firefly-Optimized CatBoost
firefly_analysis.ipynb
Evaluation and visualization of the optimized CatBoost model.
-
CVAE Architecture Definition
vae_definition.ipynb
Implements the structure of the CVAE used in the experiments. -
Mix Design Optimization
vae_optimization.ipynb
Hyperparameter tuning and model selection for CVAEs. Mixture Design Optimization using selected CVAE. -
Comparison Using Learning Curves
vae_comparison_learning_curve.ipynb
Comparative analysis of CVAE variants based on learning behavior. -
Comparison Using Loss Functions
vae_comparison_on_losses.ipynb
Performance comparison of different CVAE architectures using loss metrics.
If you use any part of this work, please cite appropriately. You may refer to the papers submitted to RILEM Youth Symposium 2025 and I-4AM’26 listed above.