改进的PSO算法在桁架结构形状优化设计中的应用

利用随机方向法初始化种群提高粒子群算法初始种群的质量．将模糊推理应用于粒子群算法的参数调整克服了人为经验设定参数的不足，种群搜索过程中嵌入Metropolis准则改善粒子群算法的鲁棒性能。将改进的粒子群算法应用于桁架结构形状优化设计中。实验仿真表明．改进后的算法具有较好的搜索性能和较高的计算精度，有望实现应用在复杂的桁架结构优化设计中．其具有重要的理论研究价值和广阔的工程应用前景.

Improved PSO Algorithm appling on shape optimization design of truss structures

Applying random direction method to produce initial PSO population improves its population quality. Fuzzy reasoning used in PSO of parameters adjustment overcomes the shortcoming relying on human experience setting parameters. The algorithm is embedded the Metropolis criteria in the process of searching that improves the robust performance of PSO. Improved PSO is applied to the tress structure shape optimization design. The simulation results show that the modified algorithm not only has the better search performance, the higher accuracy and the prospective of realizing in applying complex truss structure optimization design, but has important theoretical value and prosperous engineering application prospects.