DE/GSA: A Memetic Algorithm based on Gradient Subspace Approximation

摘要：

　　We present a new memetic algorithm for scalar optimization.The idea of the local search engine is to approximate the most greedy search direction v at a given point out of the neighborhood information that is given by the current population and to perform a line search in this direction.Two main advantages of the method are that in principle any neighboring point can be taken for the computation of v and that constraints can be incorporated directly.A first implementation of a memetic differential evolution using this techniques shows promising results.

关键词： gradient subspace approximation memetic strategy differential evolution

作者: Sergio Alvarado Oliver Schütze Carlos Segura Ricardo Landa

作者单位: Computer Science Department, Cinvestav-IPN, Av. IPN 2508, C. P. 07360, Col. San Pedro Zacatenco,Mexi Information Technology Laboratory, CINVESTAV Tamaulipas, Parque Tecnotam,km 5.5 carr. Cd. Victoria

会议类型: 国际会议

会议名称: Evolve International Conference-2014(2014年演变算法国际会议)

会议地点: 北京

会议语种:英文

页码: 1-4

在线出版日期: 2014-07-01（万方平台首次上网日期，不代表论文的发表时间）