Grey Chaotic Particle Swarm Optimization Algorithm of High Dimension Multi-Objective with Hybrid Discrete Variables

Aimed to the problems of multivariable and complicated multi-objective function and constraint conditions in mechanical optimal design, a general finding solution method (namely, grey chaotic PSO algorithm) is introduced to high dimension multi-objective hybrid discrete variables and multiobjective function is transformed into single objective function by absolute degree of grey incidences. The model and method can reasonably deal with value adopting problems of hybrid discrete variables in optimization design. Due to the morbidity problem of traditional penalty function dealing with the constraint conditions, the new method using Chaos variable to update the violating constraint solution individual is proposed for improving PSO. The new method can increase individual diversity, avoid prematurity of PSO. The program of the grey chaos multiobjective optimization with hybrid discrete variables is developed with Matlab. The practical example shows that this algorithm has no special requirements on the characteristics of optimal designing problems, which has a fairly good universal adaptability and a reliable operation of program with a strong ability of overall convergence.