在超目标优化问题中,目标之间的冲突会导致没有一个解可以同时优化所有目标,求解时存在大量非支配解.选择合适的解排序算法评估解的质量,对算法性能起着关键作用.而不同的解排序算法,在处理不同的超目标问题时有着各自的优劣.因此,本文提出一个基于投票机制和动态分配价值点的集成框架(ensemble many-objective evolutionary algorithm based on voting and dynamic value point,VDVP-EMEA),将不同解排序算法聚合在一起协同工作.首先,根据每种解排序算法的有效投票率,动态分配每个专家拥有的价值点,有效投票越多的解排序算法,相应赋予更多的价值点,反之则对价值点进行惩罚.然后使用末位淘汰制,废弃能力最差的专家的投票.其次,在环境选择过程中,使用精英选择策略,通过投票结果和价值点来定义个体适应度,适应度越大的个体越优先被选择.最后,为了测试VDVP-EMEA算法的性能,进行大量试验,将VDVP-EMEA与4种常用的单一解排序算法NSGA-III、SPEA2、BiGE、GrEA和一种先进的集成算法VMEF进行了比较.实验结果表明,VDVP-EMEA的收敛性和多样性明显优于这些算法.
We apply the filter technique to the model-based derivative-free trust region algorithm for solving nonlinear equations and nonlinear least-squares *** the work of Zhang et al.[SIAM ***.,20(2010),pp.3555-3576],we cons...
详细信息
We apply the filter technique to the model-based derivative-free trust region algorithm for solving nonlinear equations and nonlinear least-squares *** the work of Zhang et al.[SIAM ***.,20(2010),pp.3555-3576],we consider building the individual interpolation model for each component function in the nonlinear equations.A multidimensional filter,which is a list of n-tuples,is designed to potentially accept the trial point as the new iterate more *** far as we know,this is the first work to embed the filter mechanism into the model-based derivative-free *** suitable conditions,we establish the liminf-type and lim-type first-order global convergence results of the *** experiments show that the new algorithm is efficient.
暂无评论