TY - JOUR
T1 - MLPSO
T2 - Multi-Leader particle swarm optimization for multi-objective optimization problems
AU - Ibrahim, Zuwairie
AU - Lim, Kian Sheng
AU - Buyamin, Salinda
AU - Satiman, Siti Nurzulaikha
AU - Suib, Mohd Helmi
AU - Muhammad, Badaruddin
AU - Ghazali, Mohd Riduwan
AU - Mohamad, Mohd Saberi
AU - Watada, Junzo
PY - 2015
Y1 - 2015
N2 - The particle swarm optimization (PSO) algorithm, which uses the best experience of an individual and its neighborhood to find the optimum solution, has proven useful in solving various optimization problems, including multiobjective optimization (MOO) problems. In MOO problems, existing multi-objective PSO algorithms use one or two leaders to guide the movement of every particle in a search space. This study introduces the concept of multiple leaders to guide the particles in solving MOO problems. In the proposed Multi-Leader PSO (MLPSO) algorithm, the movement of a particle is determined by all leaders that dominate that particle. This concept allows for more information sharing between particles. The performance of the MLPSO is assessed by several benchmark test problems, with their convergence and diversity values are computed. Solutions with good convergence and diversity prove the superiority of the proposed algorithm over MOPSOrand algorithm.
AB - The particle swarm optimization (PSO) algorithm, which uses the best experience of an individual and its neighborhood to find the optimum solution, has proven useful in solving various optimization problems, including multiobjective optimization (MOO) problems. In MOO problems, existing multi-objective PSO algorithms use one or two leaders to guide the movement of every particle in a search space. This study introduces the concept of multiple leaders to guide the particles in solving MOO problems. In the proposed Multi-Leader PSO (MLPSO) algorithm, the movement of a particle is determined by all leaders that dominate that particle. This concept allows for more information sharing between particles. The performance of the MLPSO is assessed by several benchmark test problems, with their convergence and diversity values are computed. Solutions with good convergence and diversity prove the superiority of the proposed algorithm over MOPSOrand algorithm.
KW - Convergence
KW - Diversity
KW - Multi-objective optimization
KW - Multiple leaders
KW - Particle swarm optimization
UR - http://www.scopus.com/inward/record.url?scp=84953375264&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=84953375264&partnerID=8YFLogxK
M3 - Article
AN - SCOPUS:84953375264
SN - 1819-6608
VL - 10
SP - 17533
EP - 17538
JO - ARPN Journal of Engineering and Applied Sciences
JF - ARPN Journal of Engineering and Applied Sciences
IS - 23
ER -