Evolutionary algorithms
Swarm intelligence (Si) algorithms are special category of EAs. The essential concept of SI is the collective behavior of decentralized and self-organized swarms. There are several popular SI algorithm families that among others include particle swarm optimization (PSO), ant colony optimization (ACO), and artificial bee colony (ABC). The popularity of SI algorithms is due to the fact that they in general can handle efficiently arbitrary optimization problems. Additionally, SI algorithms as it can be found from the literature have been widely utilized to solve several problems in antennas and in wireless communications. The swarm behavior of bird flocking and fish schooling is modeled mathemati-cally by the PSO algorithm. One may find several PSO variants in the literature. The most frequently applied PSO variants include the inertia weight PSO (IWPSO) and the constriction factor PSO (CFPSO). Moreover, comprehensive learning particle swarm optimizer (CLPSO) is a PSO algorithm that has been applied to antenna design problems. The PSO algorithm is intrinsically suitable for applica-tion to real-valued problems. Thus, binary PSO (BPSO) versions should be used for solving discrete-valued problems. BPSO is one of the most popular discrete PSO algorithms. BPSO maps real values to the discrete set by using a sigmoid trans-fer function. Additionally, several new transfer functions that perform better than the original algorithm have been introduced by the authors. Furthermore, Boolean PSO is another BPSO version with main characteristic the usage of binary oper-ators for velocity and position update. Several authors have applied Boolean PSO to antenna design problems.
Evolutionary algorithms, Page 1 of 2
< Previous page Next page > /docserver/preview/fulltext/books/ew/sbew534e/SBEW534E_ch2-1.gif /docserver/preview/fulltext/books/ew/sbew534e/SBEW534E_ch2-2.gif