Application of differential evolution to optimize exponential curve model of ultimate bearing capacity of single pile

The expression of exponential curve model which is used to forecast the ultimate bearing capacity of single squeezed branch pile is a complicate and nonlinear function. The predicted results of traditional optimization methods for the parameter regression of the exponential curve model are often obtained with a great deviation because of computational complexity and artificial factors. Therefore, the adaptive accelerating differential evolution (AADE) improved by chaos initialization and adaptive adjustment of the scale factor and differential evolution strategy generation by generation is proposed, and then it is used to solve the nonlinear optimization of the model parameters and theoretical ultimate bearing capacity. Fitting calculation and analysis of measured data which is from the static load tests on squeezed branch piles is displayed. The results show that: compared with other methods, the AADE fits the measured data better and predicts the single pile ultimate bearing capacity effectively, and furthermore, it has many good properties such as fast computing speed, high accuracy, easy control variables setting, high universality, etc.