Article ID | Journal | Published Year | Pages | File Type |
---|---|---|---|---|
510266 | Computers & Structures | 2014 | 12 Pages |
•A new algorithm is proposed, that is called the Democratic Particle Swarm Optimization (DPSO).•DPSO is applied to frequency constraint structural optimization.•DPSO is capable of coping with highly non-linear, non-convex hyper-surfaces with numerous local optima.•Four numerical examples are presented to demonstrate the viability of the algorithm.
This paper represents a new algorithm for structural optimization with frequency constraints. The new algorithm is termed Democratic Particle Swarm Optimization. The emphasis is placed upon alleviating the premature convergence phenomenon which is believed to be one of flaws of the original PSO. When considering frequency constraints in a structural optimization problem, the search spaces happen to be highly non-linear and non-convex hyper-surfaces with numerous local optima and naturally the problem of premature convergence is amplified. The proposed algorithm is capable of coping with this problem. Four numerical examples are presented to demonstrate the viability of the algorithm.