Parallel Particle Swarm-Based Structural Optimization in a Distributed Grid Computing Environment
Petri Kere, Jussi Jalkanen · 48th AIAA/ASME/ASCE/AHS/ASC Structures, Structural Dynamics, and Materials Conference · 2007
The PSO algorithm is a stochastic population-based search method capable of handling multi-modal, discontinuous, and nondierentiable optimization problems. To improve the computational throughput in discrete structural optimization, an optimization system exploiting distributed grid resources has been implemented. It is shown in this paper that grid computing provides power also to structural optimization by stochastic populationbased methods. The population of individuals corresponding to the points in the design space can be increased according to the availability of processors in the grid. Any additional individuals in the population enable a higher fidelity search in the design space. Implementation of the PSO algorithm in the grid computing platform together with numerical results are reported.