摘要
进化算法适合解决多目标优化问题 ,但难以产生高维优化问题的最优解 .文中针对此问题提出了一种求解高维多目标优化问题的新进化方法 ,即多目标协调进化算法 ,主要特点是进化群体按协调模型使用偏好信息进行偏好排序 ,而不是基于 Pareto优于关系进行个体排序 .实验结果表明 ,所提出的算法是可行而有效的 ,且能在有限进化代数内收敛 .
Evolutionary Algorithms (EAs) are well suited for optimization problems with multiple objectives. Due to their inherent parallelism they are able to capture a number of solutions concurrently in a single run. Although EAs based on Pareto ranking are independent of the convexity or concavity of a tradeoff surface and do not require any preference information, the dimensionality of the search space influences their performance. Pure Pareto EAs cannot be expected to perform well on problems involving many competing objectives and may simply fail to produce satisfactory solutions due to the large dimensionality and the size of the trade-off surface. In this paper we propose a new evolutionary approach for multiobjective optimization to deal with this problem, namely MultiObjective Concordance Evolutionary Algorithm (MOCEA). Its key character lies on that evolutionary population is rather preference ranked by preference information based on concordance model than normally ranked by Pareto superior relationship. Optimization process accepts incomparability and uses an outranking relation to model the preferences of the decision-maker, which belongs to the family of outranking approaches (such as ELECTRE) introduced by Bernard Roy. This approach includes two phases: the construction of an outranking relation on the different objectives, and the exploitation of this relation in order to give an answer to the multiobjective optimization problem. Evolutionary process generates cumulative Pareto optimal solutions from which ELECTRE selects a certain subset based on the preferences of the decision-maker. The preferences are expressed through preference relationships. The advantage of the novel algorithm is that it can solve optimization problems with many objectives. Test results show that MOCEA is feasible and effective. Some experiments give a basis showing MOCEA converges to the Pareto front within a finite generation.
出处
《计算机学报》
EI
CSCD
北大核心
2001年第9期979-984,共6页
Chinese Journal of Computers
关键词
多目标协调进化算法
全局优化算法
数学模型
evolutionary algorithms, multiobjective optimization, large dimensionality, concordance model
