Convergence Analysis of a Markov Chain Monte Carlo Based Mix Design Optimization for High Compressive Strength Pervious Concrete

2018 3rd International Conference on Building Materials and Construction

土木建筑工程

Huang, Jiaqi^1 ; Jin, Lu^1

Faculty of Engineering and Mathematical Sciences, University of Western Australia, WA, Australia^1

关键词: Control parameters;    Conventional concrete;    Convergence analysis;    Convergence rates;    Effective mechanisms;    Low compressive strengths;    Markov Chain Monte-Carlo;    Optimal solutions;   

Others  :  https://iopscience.iop.org/article/10.1088/1757-899X/371/1/012020/pdf
DOI  :  10.1088/1757-899X/371/1/012020

学科分类：土木及结构工程学

来源: IOP

【 摘 要 】

Compared with conventional concrete products, pervious concrete usually features with high water permeability rate and low compressive strength due to the lack of fine aggregates. Thus the determination of optimal mix design of ingredients has been recognized as an effective mechanism to achieve the trade-off between compressive strength and permeability rate. In this paper, we proposed a Markov Chain Monte Carlo based approach to approximate the optimal mix design of pervious concrete to achieve a relatively high compressive strength while maintaining desired permeability rate. It is proved that the proposed approach effectively converges to the optimal solutions and the convergence rate and accuracy rely on a control parameter used in the proposed algorithm. A number of simulations are carried out and the results show that the proposed system converges to the optimal solutions quickly and the derived optimal mix design.