在全三维粒子模拟软件CHIPIC平台上,分别开发了粒子群及基因算法模块.以相对论返波管为例,采用三种不同类型的参数(连续参数、离散参数、混合参数),对粒子群及基因算法进行比较.优化结果表明:粒子群算法的收敛速度更快,在有限的迭代步数内得到的目标结果也更优良,总体表现优于基因算法. Particle swarm optimization and genetic algorithm are respectively developed on the full-dimensional particle simulation software CHIPIC platform.Taking the relativistic backwave tube as an example, three different types of parameters (continuous parameters, discrete parameters and mixed parameters) The results of the optimization show that the particle swarm algorithm converges faster and achieves better target results in a limited number of iteration steps, and the overall performance is better than the genetic algorithm.