Ali Reza Ghanizadeh - Department of Civil Engineering, Sirjan University of Technology, Sirjan, Iran
Nasrin Heidarabadizadeh - Department of Civil Engineering, Sirjan University of Technology, Sirjan, Iran
Meysam Bayat - Department of Civil Engineering, Najafabad Branch, Islamic Azad University, Najafabad, Iran
Vahid Khalifeh - Department of Civil Engineering, Sirjan University of Technology, Sirjan, Iran

In this study, the evolutionary polynomial regression (EPR) method has been employed to develop simple models with reasonable accuracy to predict the compressive strength and Young's modulus of the lime/cement stabilized clayey subgrade soil. For this purpose, the different specimens with the various cement and lime contents, at three moisture contents (dry side, wet side, and optimum moisture content) were fabricated and were cured for ۷, ۱۴, ۲۱, ۲۸ and, ۶۰ days to conduct the unconfined compressive strength (UCS) test. According to the test results, a dataset consisting of ۷۵ records for each additive was prepared. Results of this study show that the R۲ value of the developed model for predicting UCS of cement-stabilized clay soil is equal to ۰.۹۶ and ۰.۹۵ for training and testing sets, respectively. These two values for lime-stabilized soil are ۰.۹۱ and ۰.۸۷, respectively. Moreover, the R۲ for predicting Young's modulus of cement-stabilized clay soil is equal to ۰.۹۰ and ۰.۸۹ for training and testing set, respectively. These two values for predicting Young's modulus of lime-stabilized soil are ۰.۸۸ and ۰.۹۴, respectively. The sensitivity analysis showed that for the Portland cement stabilized clayey subgrade, the percentage of the Portland cement and moisture content are the most significant parameters for predicting the UCS and Young's modulus, respectively. In contrast, for the lime-stabilized clayey subgrade soil, the most important parameters are the moisture content and the UCS, respectively.

Stabilized clay, Portland cement and Lime, Unconfined Compressive Strength, Young' s Modulus, Evolutionary Polynomial Regression