摘要
针对如何通过附加的方法对多目标化问题进行理论分析,提出并证明了选择附加函数的3个前提条件.提出一种多目标化进化算法,根据种群中个体的多样性度量进行多目标化,并采用改进的非劣分类遗传算法对构造所得的多目标优化问题进行多目标优化.在静态和动态两种环境下进行算法性能验证,结果表明,在种群多样性保持、处理欺骗问题、动态环境下的适应能力等方面,所提算法明显优于其他同类算法.
This paper investigates the use of multi-objective evolutionary algorithms to solve single-objective optimization problems, and an instructional theory is proposed for constructing reasonable additional objective. A multi-objective optimization algorithm based on multi-objectivization is also proposed. Single-objective problems are changed to multi-objective problems based on individual diversity. The nondominated sorting genetic algorithm is used as multi-objective optimization algorithm. The computation results show that the proposed algorithm has the approving performance in maintaining population diversity, dealing schema deception problems and dynamic optimization.
出处
《控制与决策》
EI
CSCD
北大核心
2010年第9期1343-1348,共6页
Control and Decision
基金
国家自然科学基金项目(60963002)
航空科学基金项目(2008ZD56003)
江西省教育厅基金项目(GJJ08209)
关键词
多目标化
附加函数
动态环境
进化算法
移动峰问题
Multi-objectivization
Additional objective
Dynamic environment
Multi-objective optimization
Moving peaks benchmark