基于IMA-SVM的斜拉桥索力可靠度计算模型

IMA-SVM Based Cable Force Reliability Analysis Model for Cable-Stayed Bridges

针对大跨度斜拉桥结构可靠度分析时计算量大、计算效率低等问题,提出了一种基于改进蜉蝣算法与支持向量机(IMA-SVM)的可靠度计算模型。采用Logistic混沌映射初始化策略、非线性递减重力系数和子代高斯变异操作对标准蜉蝣算法进行改进,引入基准测试函数对比了单一策略和混合策略改进下蜉蝣算法的寻优性能,并将其用于求解支持向量机的最优参数和结构隐式功能函数的极值。研究结果表明,采用混合策略改进的蜉蝣算法收敛速度和精度得到大幅提升;IMA-SVM在测试集中的最大拟合误差仅为1.53%;IMA对大跨度斜拉桥索力可靠度隐式功能函数的求解效率较高,得到索力可靠度最小值为3.831,验证了上述算法对实际工程的适用性。

Aiming at the problems of large amo unts of calculation and low efficiency of rel iability calculation of long-span cable-stayed bridges,a reliability calculation method based on an improved Eph emera algorithm and support vector machine(IMA-SVM)is proposed.The standard mayfly algorithm is improved by us ing the Logistic chaos map initialization strategy,nonlinear decre asing gravity coefficient and sub-generation Gaussian mutation operation.The benchmark function is introduced to comp are the optimization performance of the Mayfly algorithm under the single strategy and the mixed strategy improvem ent,and it is used to solve the optimal param eters of the support vector machine and the extremum of the structural im plicit function.The results show that the convergence speed and accuracy of the hybrid strategy improved Ephemer a algorithm are greatly improved.The maximum error of IMA-SVM fitting the structural function in the test set is only 1.53%.IMA has a high efficiency in so lving the implicit function of cable force reliability of long-span cable-s tayed bridges,and the minimum value of cable force reliability is 3.831,which verifies the applicability of the prop osed method to practical projects.