结合遗传进化算法的全局优化思想和时间序列分析的基本理论,提出了一种新的基于遗传算法的建筑物沉降非线性动力学行为预测模型进化识别算法。设计了模型结构和参数分别进化,共同识别方案,实现对非线性时间序列分析模型结构和参数进行全局最优搜索。实例分析结果表明:该方法具有较好的预测精度和推广预测能力,为高大建筑物沉降预测提供了一个有效的分析工具。 Combining with the global optimization theory of genetic evolutionary algorithm and the basic theory of time series analysis, a new evolutionary identification algorithm of building subsidence nonlinear dynamic behavior prediction model based on genetic algorithm is proposed. The model structure and parameters are designed to evolve and identify separately, and the optimal global search for the structure and parameters of the nonlinear time series analysis model is realized. The case study shows that this method has better prediction accuracy and promotion ability, which provides an effective analysis tool for settlement prediction of tall buildings.