解WTA问题群智能优化算法的种群初始化问题研究

为提高解武器—目标分配问题(WTA)的群智能优化算法初始种群性能, 从提高初始种群多样性的角度提出了基于密码筒的种群初始化方法, 从提高初始种群解质量的角度分别提出了基于目标威胁—打击效果混合排序的种群初始化方法和装甲分队战术种群初始化方法。最后, 通过一系列不同规模WTA算例的仿真实验对初始化方法进行评估, 结果表明三种种群初始化方法均达到了设计要求, 且各具优势。

Research on population initialization forswarm intelligence optimization solving WTA

In order to improve the performance of initial population for swarm intelligence optimization solving weapon-target assignment, this paper established the population initialization method based on cryptex to improve population diversity of initial population, and established the population initialization method based on threat-capability matrix sorting and armored element tactic population initialization method to improve solution performance of initial population. Finally, several simulation experiments with WTA examples of a series of scales were carried on to evaluate initialization methods. The results indicate that three methods have reached the design requirements, and have respective advantages.