作者: Juan Pablo Serrano-Rubio , Arturo Hernández-Aguirre , Rafael Herrera-Guzmán
DOI: 10.1007/S00500-016-2461-Y
关键词: Beam stack search 、 Guided Local Search 、 Beam search 、 Operator (computer programming) 、 Binary search algorithm 、 Mathematics 、 Mathematical optimization 、 Best-first search 、 Evolutionary algorithm 、 Search algorithm
摘要: This paper introduces an evolutionary algorithm which uses reflections and spherical inversions for global continuous optimization. Two new geometric search operators are included in the design of algorithm: inversion operator reflection operator. The computes inverse points with respect to hyperspheres, redistributes individuals on space fitness function. nonlinear nature furnishes more “aggressive” exploitation capabilities algorithm. performance is analyzed through a benchmark 28 functions. Statistical tests show competitive comparison current leading (geometric) algorithms such as particle swarm optimization four differential evolution strategies.