Identification of multiple alternative solutions in structural model updating using multiple subswarm particle swarm optimization

Abstract

The global optimal solution of structural model updating may not represent the real physical structure owing to modelling simplifications, measurement errors and uncertainties. This study proposes a multiple subswarm particle swarm optimization (MSPSO) to identify multiple alternative solutions in structural model updating. The MSPSO introduces two additional operators, a double angle operator and a multiple subswarm operator, into the standard particle swarm optimization (PSO). The former aims to identify the search range of possible solutions and the latter facilitates search efficiency after the search range has been located. Numerical and engineering applications, including two benchmark functions, a space truss structure and a large-span arch bridge, are used to examine the feasibility of the proposed method. The results indicate that MSPSO can identify multiple alternative solutions effectively while PSO only finds the global optimal solution, and these solutions provide more options for the selection of the best physical representation of the structure.