ABSTRACT
This paper examines a new approach of model updating in bridge structure based on the analytic hierarchy process (AHP) and Kriging meta-model, which overcomes the difficulties in determining weight coefficients for model updating process. The Kriging meta-model performs as a surrogate model for predicting the analytical responses to avoid complex computation using the finite element (FE) model. To deal with multiple objectives and enable to make decisions according to different demands on each evaluation index, the AHP is implemented to determine the weight coefficients. A plane truss bridge is applied to verify the proposed approach. The results show that the integration of AHP, differential evolution (DE) algorithm and Kriging meta-model could comprehensively update the initial model and strongly reduce the computational cost. The investigation results have revealed that the hybrid method could serve for model updating more efficiently and effectively and prove its potentiality for solving practical problems in real engineering structures.
