给出了动态多目标优化问题的一种新解法.首先对时间变量进行了等区间离散化,在得到的子区间(称为环境)上定义了种群的静态序值方差和静态密度方差.然后把动态多目标优化问题近似地转化成了若干个两个目标的静态优化问题.在给出的一种能自动检测环境变化的应答算子下,提出了一种动态多目标进化算法,同时证明了算法的收敛性.计算机仿真表明新算法对动态多目标优化问题是有效的.
A method for dynamic multi-objective optimization problems (DMOPs) is given.Ftrst, we divide the time period into several equal subpefiods. In each subpefiod (termed as environment ), the static rank variance and the static density variance of the population are defined,thus the DMOPs is transformed into several bi-objective static optimization problems by using the static rank variance and the static density variance. Then, based on a new mutation operator which can automatically check out the environment variation, a dynamic multi-objective evolutionary algorithm is proposed and the convergence analysis of the algorithm is presented. Finally the numerical results demonstrate the effectiveness of the new algorithm.