IGLO algorithm-based large-scale structural optimization with mixed-discrete design variables

摘要

The efficient stochastic-type Integral Global/Local Optimization (IGLO) algorithm-based structural design methodology advanced for the continuous design variable case is extended and applied to constrained structural optimization problems of mixed discrete/continuous design variable-type in a massively parallel environment. ^The Feasible Design Set Only (FDSO), as well as the improved constraint treatment schemes of Quasi-Linear Design Scaling (OLDS) and the Modified Exterior Penalty Function (MEPF) are formulated and implemented. ^(AIAA)