中文摘要：

动态多目标优化是进化计算领域一个新兴的研究方向．文中给出了定义在离散时间空间上、决策变量的维数随时间（环境）可发生变化的一类动态多目标优化问题（DMOP）的新方法．该方法首先把DMOP转化成了一系列同类静态约束优化问题，然后在一种环境变化判断规则下提出了解DMOP的一种新动态多目标进化算法（DMEA）．数值实验表明新算法对DMOP最优解具有较好的跟踪能力，并且能有效的获得DMOP在不同环境下数量较多、质量较好且分布均匀的Pareto最优解．

英文摘要：

Dynamic multi-objective optimization is a new area of evolutionary computation. A method solving a class of dynamic multi-objective optimization problem （DMOP） which defined in the discrete time space and the dimension of decision variable changing with time （environment） is given. First, the DMOP is transformed into a series of homogeneous static constraint optimization problems. Then, a new dynamic multi-objective evolutionary algorithm （DMEA） is proposed based on a role which can automatically check out the environment variation. At last, the experimental results and comparison illustrate the proposed algorithm can track the Pareto optimal solutions of DMOP and find a sufficient number of uniformly distributed Pareto optimal solutions in difference environment.