Improvement for an optimization algorithm and its application in structural model updating

Yonghui An

doi:10.1007/s13349-025-00915-z

2025-02-15

PRE

DOI：10.1007/s13349-025-00915-z

原文链接

原文求助

摘要

En 中文

Model updating is one of the important components of structural health monitoring. Artificial fish swarm algorithm (AFSA) is an effective optimization algorithm and can be used in model updating and many research fields. However, its efficiency still requires improvement. A series of improvements for the AFSA is proposed, i.e., accelerating the preying behavior, and setting swimming direction in the previous iteration as the prior direction in the next iteration. After an iteration, the direction of the maximum concentration point within visual distance is selected to swim towards and this direction information is saved; in the preying behavior of the next iteration, the saved direction is considered as the prior direction to prey; if the food concentration in the saved direction is higher than current position, the fish will swim a step in this direction; otherwise, it will prey randomly again. The De Jong Function is used to validate the proposed method and results show that the improvement enhances convergence performance and higher efficiency. Moreover, its application in the finite element model updating of a long-span prestressed concrete continuous rigid frame bridge is investigated; seventeen parameters of the bridge are selected to update its numerical model, and results show that the proposed improvement is successfully applied to model updating. Compared with the AFSA without the proposed improvement, it is more efficient and it saves 25.314% and 32.742% in computational time for the present two examples, respectively. The proposed improvement can be used in various optimization problems of different fields.

Keyword:

Artificial fish swarm algorithm

Model updating

Rigid frame bridge

Structural health monitoring

Optimization