基于代理模型的QPSO算法及结构优化应用

针对复杂结构的工程优化通常涉及到高度非线性的问题，采用基于迭代的传统优化算法可能无法获得全局最优解。在此研究背景下，文章引入量子粒子群优化算法(Quantum Particle Swarm Optimization, QPSO)，提出了基于克里金（Kriging）代理模型的QPSO算法，通过构建环形桁架可展天线动力学性能的代理模型，对环形桁架可展天线结构参数进行了优化设计的应用研究。研究结果表明，基于Kriging代理模型的QPSO算法计算效率较高，环形桁架可展天线的基频及最大冲击响应面非线性特征显著，并通过将计算所得到的最优解与有限元解进行对比，验证了此方法的计算精度。由此可见，此方法的研究可为复杂结构的优化设计、研制提供一定的参考。

Quantum Particle Swarms Optimizations Algorithm Based on Surrogate Model for the Application of Structural Optimization

The traditional optimization algorithms might fail to get the global optimal solution for high non-linear characteristics in engineering optimization of multi-structures. So the Quantum Particle Swarm Optimization (QPSO) algorithm was induced and the QPSO algorithm based on surrogated model was studied. The applications to the structural parameter optimization of the ring truss deployable antennas were carried out by the surrogated model of dynamic characters for ring truss deployable antennas. The results show that the computation efficiency of QPSO algorithm based on Kriging model is acceptable. The non-linear characteristics of the 1^st natural frequency and maximum impulse response of the ring truss deployable antennas are remarkable. Then the computation accuracy is validated by comparing the solutions between the optimal result and the finite element result. It′s proved that QPSO algorithm based on Kriging model can be used to optimize the complicated structures in engineering design.