Study on Optimal Setup of the ANN-Aided Aimed Multilevel Sampling Optimization Method

Abstract

This paper deals with the application of Aimed multilevel sampling metaheuristic optimization method supported by an artificial neural network. The main aim is to study the optimal parameter settings of this method for damage identification on a two-span steel truss using a limited number of simulations. The convergence rate is tested as a function of the number of simulations and the choice of the parameter that controls the resizing of the design space. The results are summarized and discussed with respect to their practical applicability in bridge damage identification using structural health monitoring data.