Multiobjective two-stage optimization of a plate structure using a population-based incremental learning method

Abstract

This paper proposes a multiobjective two-stage optimization and investigates on the performance of a meta-heuristic namely multi-objective population based incremental learning (MOPBIL) for solving two-stage optimization of a plate structure. A design process consists of two stages as (1) topology optimization and (2) shape and sizing optimization. In the topology optimization process, the ground element filtering (GEF) is used to suppress checkerboard patterns on a structural topology and to reduce a number of design variables. MOPBIL with various sets of optimization parameter settings is applied to tackle both design stages. Comparative results based on the hypervolume indicator reveal that changes in parameter setting slightly affect MOPBIL search performance. Structural topology and structural shape and sizing results obtained from using MOPBIL are illustrated and compared. It is illustrated that the use of MOPBIL to find optimal shape and sizing of plate structure is effective and practical.