近年来，多目标优化与决策问题求解已成为演化计算的一个重要研究方向。为使演化算法的种群解 能尽快收敛并均匀分布于多目标问题的非劣最优域，多目标演化算法的研究热点集中在基于Pareto最优概念的 种群个体的比较与排序、适应值賦值与小生境技术等方面。介绍了多目标优化与决策技术的发展历史与分类方 法，分析了基于Pareto最优概念与不基于Pareto最优概念两大类的多目标演化算法，并详细比较与分析了几种 典型多目标演化算法。其次，论述了与多目标演化算法研究紧密相关的一些问题，如多目标问题解的性质，测 试函数集设计，算法性能评估技术，算法收敛性，并行实现以及实际多目标优化问题的处理等。

Multi-objective optimization (MOO) and decision-making (DM) has become an important research area of evolutionary computations in recent years. The researches on multi-objective evolutionary algorithms (MOEA) focus mainly on the Pareto-based comparison and ordering of individuals, fitness assignment and Riching techniques, etc., so that the population can converge and uniformly distribute in the Pareto front. This paper presents an introduction to the history and classification of multi-objective optimization and decision-making techniques, analyzes both the Pareto-based and non-Pareto-based evolutionary algorithms, and,particularly,the five well-known MOEAs. Some problems related to the researches on MOEAs are addressed in details, such as the characteristics of Pareto front, the test suite and performance evaluation of MOEAs, the MOEA convergence analysis, the MOEA parallelization, and the disposal of real world MOO problems.