Randomization in particle swarm optimization for global search ability

Adaptive division of labor particle swarm optimization

“…Wu, et al (2014) introduced a superior solution-guided PSO framework in which each particle can comprehensively learn from a collection of superior solutions. Zhou, et al (2011) introduced random position PSO (RPPSO) in which a random particle is used to guide a swarm if a randomly generated number is smaller than the proposed probability Chen, et al (2013) incorporated an aging mechanism into PSO and proposed a PSO with an aging leader and challengers (ALC-PSO). When the leader of the ALC-PSO swarm is no longer effective in improving the population, its leading power is gradually deteriorated and eventually replaced with a new emerging particle that aims to challenge and claim the leadership.…”

Particle swarm optimization with dual-level task allocation

“…The majority of the tested PSO variants undergo different degrees of performance degradation while addressing rotated problems because the rotating operation introduces a non-separable characteristic into this problem category. (Zhan et al, 2009) Fully connected ω : 0:9 À 0:4, c 1 þ c 2 : ½3:0; 4:0, δ ¼ ½0:05; 0:1, σmax ¼ 1:0, σ min ¼ 0:1 FLPSO-QIW (Tang et al, 2011) Comprehensive learning ω : 0:9 À 0:2, c 1 : 2 À 1:5, c 2 : 1 À 1:5, m ¼ 1, P i ¼ ½0:1; 1, K 1 ¼ 0:1, K 2 ¼ 0:001, σ 1 ¼ 1, σ 2 ¼ 0 FlexiPSO (Kathrada, 2009) Fully connected and local ring ω : 0:5 À 0:0, c 1 ; c 2 ; c 3 : ½0:0; 2:0, ε ¼ 0:1, α ¼ 0:01% FPSO (Montes de Oca et al, 2009b) Time-varying χ ¼ 0:729, ∑c i ¼ 4:1 OLPSO-L (Zhan et al, 2011) Orthogonal learning ω : 0:9 À 0:4, c ¼ 2:0, G ¼ 5 PAE-QPSO (Fu et al, 2012) Fully connected μ ¼ ½0; 1, β : 1:0 À 0:5, RPPSO (Zhou et al, 2011) Random ω : 0:9 À 0:4, c large ¼ 6, c small ¼ 3 MoPSO (Beheshti et al, 2013) Fully connected No parameters are involved PSODDS (Jin et al, 2013) Fully connected χ ¼ 0:7298, c 1 ¼ c 2 ¼ 2:05 PSO-DLTA Fully connected and local ring ω : 0:9 À 0:4, c 1 ¼ c 2 ¼ 2:0, z¼ 8…”

“…The selection of strategies for each particle is based on a ratio that is derived from a self-adaptively improved probability model. Zhou et al (2011) introduced random position PSO (RPPSO), wherein a random particle is used to guide the swarm if a randomly generated number is smaller than the proposed probability. Ho et al (2008) developed the orthogonal PSO (OPSO) using the orthogonal experiment design (OED) technique (Montgomery, 1991).…”

Buffer Insertion for Delay Minimization using An Improved PSO Algorithm

“…The acceleration factors c 1 and c 2 indicate the relative attraction toward pbest and gbest respectively and also rand 1 and rand 2 are random numbers uniformly distributed between zero and one [13]. Inertia weight parameter w controls the trade-off between the global search ability and local search ability during the optimization process [23]. In order to avoid premature convergence, PSO utilizes a distinctive feature of controlling a balance between global and local exploration of the search space which prevents from being stacked to local minima [13].…”