AGENT-BASED HYBRID METAHEURISTIC FOR NUMERICAL MODEL UPDATE

This framework aims to provide flexible nondeterministic strategies to guide the numerical model updating process.

Despite advances in instrumentation and measurement techniques, numerical models are still crucial to simulate and estimate dynamic responses. However, such models must be updated for consistency between simulation and reality. Thus, ,this chapter proposes a hybrid metaheuristic structure based on agents and an attribute classification algorithm for updating finite element models (FE).

This framework aims to provide flexible nondeterministic strategies to guide the update process, ranging from simple local search procedures to complex learning processes. The case study are presented a free-free aluminium beam tested under laboratory conditions. The upgrade aimed to optimize the stiffness matrix while keeping the mass matrix unchanged.

As for the updated numerical vibration modes, the Modal Assurance Criteria (MAC) values slightly decreased in both cases but within the acceptable MAC values (above 0.9), thus showing good consistency with the experimental vibration modes.