HYSTERESIS AND VISCOUS DAMPER PLACEMENT OF ELASTO-PLASTIC MULTI-STORY STRUCTURES USING SEQUENTIAL APPROXIMATE MULTI-OBJECTIVE OPTIMIZATION

Abstract

This paper proposes a sequential approximate multi-objective optimization method and its application to the hysteresis and viscous damper placement problem of elasto-plastic multi-story structures. The objective and constraint functions are approximated by using machine-learning techniques. Search for the optimal solutions, or the Pareto frontier, is conducted mostly in the approximated response space and only promising solutions are forwarded to true function evaluations with costly time-history analysis. The validity of the method is shown in a 10-story lumped mass shear spring model subjected to earthquake input.