论文部分内容阅读
网格策略在对多目标粒子群优化算法档案的多样性分析时具有简单快捷的特点,但由于其分辨率的限制,无法对含有相同粒子数量的超体进行多样性判断,对此,提出了动态网格的策略,通过目标空间的动态划分、超体的动态调节等手段,对多样性相同的超体进行密度排序,方便算法对档案进行管理和选择领导粒子,同时利用了以往被忽视的档案解,对其进行基因交换,提高算法的收敛速度.通过DTLZ系列函数的验证,表明了算法在高维多目标优化中仍具有良好的多样性和更快的收敛速度,能有效解决高维多目标问题.
Grid strategy has the characteristics of simple and quick when analyzing the diversity of multi-objective particle swarm optimization algorithm files. However, due to the limitation of resolution, it is impossible to judge the diversity of hypersensors with the same number of particles. In this regard, Dynamic grid strategy, through the dynamic division of the target space, the dynamic adjustment of the super body and other means, the density of the same diversity of the same sort order to facilitate the management of the file algorithm and the selection of leading particles, taking advantage of the past has been neglected File solution, gene exchange and improve the convergence speed of the algorithm.According to the DTLZ series of functions, it shows that the algorithm still has a good diversity and faster convergence speed in high-dimensional multi-objective optimization, which can effectively solve the high dimensional Multi-goal problem.