Journal Search Engine
Search Advanced Search Adode Reader(link)
Download PDF Export Citaion korean bibliography PMC previewer
ISSN : 1598-7248 (Print)
ISSN : 2234-6473 (Online)
Industrial Engineering & Management Systems Vol.6 No.1 pp.64-71

An Improvement of Particle Swarm Optimization with A Neighborhood Search Algorithm

Fumihiko Yano, 1Tsutomu Shohdohji, 2Yoshiaki Toyoda
J. F. Oberlin University 3758 Tokiwa-Machi, Machida, Tokyo 194-0294, Japan
1Nippon Institute of Technology 4-1 Gakuendai, Miyashiro-Machi, Saitama 345-8501, Japan,

2Aoyama Gakuin Unversity 5-10-1 Fuchinobe, Sagamihara, Kanagawa 229-8558, Japan
Received Date, September 2006; Accepted Date, January 2007


J. Kennedy and R. Eberhart first introduced the concept called as Particle Swarm Optimization (PSO). They applied it to optimize continuous nonlinear functions and demonstrated the effectiveness of the algorithm. Since then a considerable number of researchers have attempted to apply this concept to a variety of optimization problems and obtained reasonable results. In PSO, individuals communicate and exchange simple information with each other. The information among individuals is communicated in the swarm and the information between individuals and their swarm is also shared. Finally, the swarm approaches the optimal behavior. It is reported that reasonable approximate solutions of various types of test functions are obtained by employing PSO. However, if more precise solutions are required, additional algorithms and/or hybrid algorithms would be necessary. For example, the heading vector of the swarm can be slightly adjusted under some conditions. In this paper, we propose a hybrid algorithm to obtain more precise solutions. In the algorithm, when a better solution in the swarm is found, the neighborhood of a certain distance from the solution is searched. Then, the algorithm returns to the original PSO search. By this hybrid method, we can obtain considerably better solutions in less iterations than by the standard PSO method.