Research on Cable Force Optimization of Cable-stayed Bridges Based on Improved Seagull Algorithm and Support Vector Machine

Aiming at the cable force optimization problem of long-span cable-stayed bridge, a cable force optimization method based on hybrid strategy improved seagull optimization algorithm and support vector machine is proposed. The standard seagull optimization algorithm is improved by integrating refraction reverse learning, multi-directional spiral attack and nonlinear convergence strategies. The penalty factor and kernel function parameters of support vector machine（SVM） are optimized by improved seagull optimization algorithm（ISOA）, and a prediction model of cable force combination structure response of long-span cable-stayed bridge is constructed. The cable force optimization model based on ISOA-SVM is designed and the cable force of stay cable under the control of girder alignment is optimized. The results show that the hybrid strategy significantly improves the convergence speed and convergence accuracy of seagull optimization algorithm. The SVM optimized by ISOA parameters has a good learning generalization ability for data samples. The average relative error of its test set is only 1.08%, and the root mean square error is only 0.012 2. The cable force combination based on ISOA-SVM optimization effectively improves the alignment and internal force of the main beam. The peak value reduction of the vertical deflection of the main beam is 36%, and the peak value reduction of the bending stress is 11.94%, which verifies the effectiveness of the cable force optimization method.