Machine-Learning-Aided Concrete Mixture Optimization

It is labour and resource intensive to optimize mixture proportions of concrete with multiple competing objectives. The aim of this thesis is to solve the multi-objective mixture optimization problem. Several imputation methods were firstly developed to replace missing values in the datasets. Then, machine learning models were used to predict properties of concrete to establish objective functions. Finally, multi-objective mixture optimization was conducted using metaheuristic optimization algorithms. This research assists in paving the way for intelligent construction in civil engineering.