Modified Particle Swarm Optimization For Structure Design Of Autobody

The structure optimization design of auto-body is a complex design optimization problem,in which the crash safety,vibration,noise and comfort of a car are all critical performance to consider.Traditional optimization algorithms cannot find a feasible solution for such problem,because of the large number of design variables and strong nonlinearity.In this dissertation,a particle swram optimization bsed approach is proposed to meet these challenges.Specifically,the spatial position distribution method of the particles,the diversity maintaining mechanism of a particle swarm,and the design sapce identification technology are developed.The main research tasks and conclusions are summarized as follows:(1)Research on the spatial position distribution method of the particles based on design of experimental and data mining technologyThe particle swarm optimization algorithm is prone to premature convergence because of the irrational population distribution.Based on design of experimental and data mining technology,the spatial position distribution method of the particles is studied and the particles distribution control method is proposed to improve the convergence of particle swarm optimization.The advantages of the proposed algorithm are verified by the analysis of standard test functions with several similar improved algorithms.(2)Research on the optimizing control method of the particle swarm optimization based on diversity maintaining mechanismsSince the PSO optimization process is prone to converging to a local optimal solution,the diversity maintaining mechanism is studied and the adaptive velocity reset mechanism is proposed to avoid particle swarm converge into local optimum,so that a global solution can be found and the optimization capability can be improved.The effectiveness of the improved algorithm is verified based on the results of mathematical examples and a practical engineering case.(3)Research on the particle swarm optimization for the constrained problem using the optimization domain identification techniqueSince constrained problems are complex and hard to find the global optimal solution,the proper constraint handling method is studied and chosen for the boundary constraints and function constraints.The modified particle swarm optimization in consideration of the features of constrained problem is proposed based on the sensitivity analysis method and data mining technique to improve the efficiency and accuracy of constrained optimization process.(4)Research on the application of auto-body structure optimization based on the modified particle swarm algorithmThe auto-body stucture optimization methodology is proposed based on the spatial position distribution method of the particles,the diversity maintaining mechanism of a particle swarm and the design sapce identification technology.The overall framework of auto-body structure optimization is established and verified through a practical engineering case.