一种改进的微分进化算法及其在多目标访问任务中的应用.

An optimal problem of orbit maneuver when multiple targets need to be visited in a described period of time is modeled and solved in this paper. Firstly, the problem description and modeling process considering multiple constraints and performance indexes can reduce the searching space and improve the efficiency. Then focusing on the drawbacks of the classical differential evolution algorithm, an improved algorithm is proposed and applied on the multiple visiting target problem. The traditional algorithm is improved by designing the double self-adaption control parameters and introducing a random mutant to improve the efficiency of optimizing and make the algorithm can jump out when trapped into the local optimum. Specific simulation verifies that the SA-DE-RM (rand) algorithm has a better performance on average value, best result and mean square error and it is feasible and effective. [ABSTRACT FROM AUTHOR]