中文摘要：

针对深空探测时间长、成本高的特点,文中以多任务深空探测为背景,建立了包含多模型、多约束、多变量、多阶段的探测轨道设计模型。通过融合推力工作模式、分段设计目标函数等策略,克服了模型的内点约束限制,降低了优化模型的复杂度。为了提高优化速度、提升优化精度,结合梯度搜索和粒子群算法的特点,研究提出了一种梯度混合粒子群（GHPSO）算法。将该算法应用到多任务星探测轨道设计模型上,得到了发动机的工作时序,并横向对比了该文算法与传统粒子群算法（PSO）和遗传算法（GA）的优化性能。仿真结果表明,文中提出的算法搜索速度快,第1设计阶段GHPSO相对GA提高61.96%,相对PSO提高47.85%,第2设计阶段GHPSO相对GA提高61.87%,相对PSO提高43.66%;精度最高,第1设计阶段GHPSO相对GA平均精度提高了2.86%,最高精度提高了1.24%,相对PSO平均精度提高了4.19%,最高精度提高了3.97%,第2设计阶段GHPSO相对GA平均精度提高了3.33%,最高精度提高了1.63%,相对PSO平均精度提高了4.72%,最高精度提高了3.02%,适用于轨道优化设计类的非线性、多约束全局优化问题。

英文摘要：

Based on long time and high cost features of deep space exploration,an optimization model,containing multiple models,constraints,variables and multi-stage,was established.By integrating thrust mode and step-by-step design of objective function,the complexity of optimization model was reduced.In order to improve the speed and accuracy of optimization algorithms,a Gradient Hybrid Particle Swarm Optimization（GHPSO） algorithm was presented,which combined the characteristic of gradient search and Particle Swarm Optimization algorithm.This algorithm was applied to the trajectory optimization for multitask explorations to obtain engine＇s timing.Then the performance of GHPSO was compared with traditional particle swarm optimization（PSO） algorithm and genetic algorithm（GA）.The results of simulation show that the GHPSO has faster convergence and higher accuracy.In the first stage of simulation,the convergence speed of GHPSO was 61.96% faster than GA and 47.85% than PSO,meanwhile,the average accuracy of GHPSO was 2.86% higher than GA and 4.19% than PSO,and the highest precision of GHPSO was 1.24% higher than GA and 1.63% than PSO.In the second stage,the convergence speed of GHPSO was 61.87% faster than GA and 43.66% than PSO,meanwhile,the average accuracy of GHPSO was 3.33% higher than GA and 4.72% than PSO,and the highest precision of GHPSO was 1.63% higher than GA and 3.02% than PSO.The result shows that the GHPSO can be applied to trajectory optimization design as well other nonlinear constrained global optimization problems.