沉降预测广义Usher模型参数确定的优化算法

为提高软土路基沉降预测精度,将用于资源预测的广义Usher模型引入软土路基沉降预测中;基于最小二乘原理,利用差分进化算法优化求解模型结构参数。分析表明,广义Usher模型包含了目前用于沉降预测的指数曲线模型、Logistic模型、Gompertz模型、Bertalanffy模型、Weibull模型、Usher模型共6种沉降预测模型,该模型具有更强大的可塑性和适应性。实例计算结果表明,经差分进化算法优化的广义Usher模型用于沉降预测效果良好,比Logistic模型、Gompertz模型、Bertalanffy模型、Usher模型这几种沉降预测模型具有更高的预测精度,可供工程应用参考。

Optimization algorithms for parameters of generalized Usher model of settlement prediction

To improve the prediction accuracy of settlement of soft clay roadbed,the generalized Usher model originally for resources forecast was introduced in the settlement prediction.Based on the least square principle,the differential evolution algorithms were employed to solve the structural parameters of the model.The generalized Usher model includes the current six kinds of models for settlement prediction: the exponential curve model,the Logistic model,the Gompertz model,the Bertalanffy model,the Weibull model and the Usher model,indicating that it has more powerful flexibility and adaptability.The calculated results of examples show that the generalized Usher model optimized by the differential evolution algorithms has high prediction accuracy and can be applied in actual projects.