.png){kind=icon}
.png){kind=icon}
Return
Intelligent mixture optimization for stabilized soil containing solid waste based on machine learning and evolutionary algorithms
.png){kind=icon}
PRE
AI
DOI：10.1016/j.conbuildmat.2024.137794.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
The usage of industrial solid waste to improve soil for road materials has attracted widespread attention. In addition to mechanical performance, economic and environmental factors gain increasing concern during road construction. This study proposes an intelligent mixture design method based on machine learning and multiobjective evolutionary algorithms. Six machine learning models for predicting California bearing ratio value of stabilized soil were developed and evaluated utilizing a dataset containing 403 samples. With the best prediction model, three multi-objective evolutionary algorithms were adopted to optimize the three objective functions including CBR, material cost, and carbon emission. The multi-objective optimization model established by combining Extreme Gradient Boosting and Non-dominated Sorting Genetic Algorithm-II successfully found the Pareto front for the three-objective optimization problem under various decision preferences. Eventually, the weights of each optimization objective were determined with subjective-objective combination assignment, which was coupled with the Technique for Order Preference by Similarity to Ideal Solution method to determine the optimal solution. The results suggest the importance of determining the optimal solution based on the demand of the decision maker and data information. The proposed framework can comprehensively consider the mechanical, economic, and environmental objectives, achieving the multi-objective optimization design of stabilized soil. This study offers practical value for the application of solid waste stabilized soils in road construction and provides an alternative towards the intelligent utilization of solid waste material.
Keywords:
Multi-objective optimization
Machine learning
Evolutionary algorithms
Solid waste
Stabilized soil
Journal
IF:
8
Papers:
4.5W
Citations:
27.9W