High performance prediction of soil compaction parameters using multi expression programming

Created by W.Langdon from gp-bibliography.bib Revision:1.7630

@Article{WANG:2020:enggeo,

• author = "Han-Lin Wang and Zhen-Yu Yin",
• title = "High performance prediction of soil compaction parameters using multi expression programming",
• journal = "Engineering Geology",
• volume = "276",
• pages = "105758",
• year = "2020",
• ISSN = "0013-7952",
• DOI = "doi:10.1016/j.enggeo.2020.105758",
• URL = "http://www.sciencedirect.com/science/article/pii/S0013795220305810",
• keywords = "genetic algorithms, genetic programming, Soil compaction, Optimum water content, Maximum dry density, Atterberg limits, Grain size distribution",
• abstract = "Previous prediction models for soil compaction parameters were developed using limited data of specific soils and their accuracy also needs to be improved. This study presents the development of a new prediction model for the soil compaction parameters (i.e. optimum water content and maximum dry density) using the multi expression programming (MEP). Numerous soil compaction tests with a wide range of soil classifications and compaction energies are first collected to form a large database. Then, the optimal setting of the MEP code parameters is investigated and determined. The explicit formulations for the two key compaction parameters are finally proposed. The validity and the sensitivity analysis of the model are conducted. The results show that the proposed model enables to predict the soil compaction parameters for all kinds of soils in the database with high accuracy. The monotonicity analysis of the predicted compaction parameters with each input property (four physical properties of soil and one compaction energy) verifies the correctness and the validity of proposed model, showing consistency with the monotonicity concerning the actual data in the database. From the sensitivity analysis about the relevance of each input property on the predicted compaction parameters, it is indicated that the plastic limit and the fines content have more significant influences on the prediction results, while the effect of the liquid limit is the least pronounced",

}

Genetic Programming entries for Han-Lin Wang Zhen-Yu Yin

Citations