Identification of strategy parameters for particle swarm optimizer through Taguchi method

摘要: Particle swarm optimization (PSO), like other evolutionary algorithms is a population-based stochastic algorithm inspired from the metaphor of social interaction in birds, insects, wasps, etc. It has been used for finding promising solutions complex search space through particles swarm. well recognized fact that performance to great extent depends on choice appropriate strategy/operating parameters population size, crossover rate, mutation operator, Generally, these are selected hit and trial process, which very unsystematic requires rigorous experimentation. This paper proposes systematic based Taguchi method reasoning scheme rapidly identifying strategy PSO algorithm. The robust design approach using fractional factorial study large number with small experiments. Computer simulations have performed two benchmark functions—Rosenbrock function Griewank function—to validate approach.