摘 要：本文针对以结构动力响应为约束，最小重量为目标的桁架拓扑优化问题，提出了一种将微粒群算法和优化准则法结合的混合PSO算法。利用优化准则法的迭代关系找出群体中适应度最好的微粒，将其作为特殊微粒，其他微粒的寻优采用PSO的基本进化规则，位移响应约束利用特殊微粒的灵敏度信息近似计算。算例的计算结果表明，混合PSO算法适用于受简谐荷载以及脉冲荷载作用桁架结构的拓扑优化。混合PSO的计算效率比PSO算法高，其优化效果比优化准则法好。

Abstract: A hybrid algorithm is proposed for the dynamic topology optimization of truss structures. In optimization analyses the dynamic displacement responses of nodes are constrainted and the minimal weight of the truss structure is as the objective function. A particle which has the best fitness from the swarm is chosen as the special particle, and it moves by means of iteration expressions in the optimality criteria methods. The other particles comply to the evolution rules of PSO algorithm, and their displacement constraints are approximately calculated by using the dynamic sensitivity of the special particle. The optimization results of a computational example show that the hybrid PSO algorithm is suitable for the topology optimization of truss structures under the harmonic load and impulsive load. The optimization efficiency of hybrid PSO algorithm is higher than PSO algorithm, and the optimization precision of hybrid PSO algorithm is better than the optimality criteria method.