基于RAGA的指数曲线模型预测基桩承载力

预测挤扩支盘桩单桩极限承载力指数曲线模型的函数表达式是一个超定非线性方程。采用传统的最小二乘法对该模型参数进行回归处理时，往往由于计算复杂和人为因素的影响使预测结果存在较大的误差。为此，用实数编码加速遗传算法（RAGA），对指数曲线模型的参数和理论极限承载力进行优化，并基于MATLAB环境下编写了模型的计算程序。结合工程实例，应用指数曲线模型对挤扩支盘桩静载荷试验实测数据进行了拟合分析。结果表明，基于RAGA的指数曲线优化模型能够更好地拟合实测数据和有效地预测单桩极限承载力，且基于RAGA的指数曲线优化模型具有求解速度快、计算精度和自动化程度高、人为干扰因素小、通用性强等优点。

Application of RAGA-based exponential curve model in prediction of pile limit bearing capacity

The equation of exponential curve model, which used to predict single pile limit bearing capacity, is super-set and nonlinear. Traditional optimization methods including the least squares method on the parameters regression of the exponential curve model often make the predicted results with a great deviation because of computational complexity and artificial factors. Therefore, the real coding based accelerating genetic algorithm (RAGA) is used to optimize the parameters and theoretical limit bearing capacity of exponential curve model. The RAGA solution program is compiled in the software MATLAB environment. Then some contrastive model applications to branch-piles is given. The results show that the RAGA based exponential curve model can better fit the measured data and effectively predict the limit bearing capacity, and RAGA is a high effective algorithm with many good properties such as high efficiency, high precision, fast computing speed, small artificial factors, etc.