Table 2.

List of used PSO variants.

PSO algorithm	Abbreviation	Changes to flowchart in Figure 1	Hyperparameter values for test performance	Hyperparameter dependence
RPSO [35]	M1	A particle $i\leftarrow t\mathrm{mod}P$ (where t is the current generation and P is the population size) is randomly sampled in the domain before Step 5	Particles per generation to sample: 1	Low
PSO with constriction factor [34]	M2	–	–	Very low
PSO + Genetic Algorithm [43]	M3	Genetic Algorithm is performed before Step 5	$pc=0.8$	High
			$pm=0.005$
			GA performance per generation: 1
PSO + Cuckoo Search (CS) [44]	M4	Position (in Step 4) is calculated based on the following equation: ${\mathbf{x}}_{t+1}^i=R_t^i{\mathbf{x}}_t^{i,PSO}+\left(1-R_t^i\right){\mathbf{x}}_t^{i,CS}$	CS performance per generation: 1	Low
		where ${\mathbf{x}}_t^{i,PSO}$ is calculated using (2), $R_t^i$ is a random number and ${\mathbf{x}}_t^{i,CS}$ is calculated using Cuckoo-Search related equations (in [44])
PSO + Firefly Algorithm (e.g. [45])	M5	The population is split into two subpopulations SP1 and PS2. Firefly Algorithm is performed on PS1 and PSO equations (Step 4) on SP2	${\beta }_0=\gamma =1$ $\alpha =0.1$ ${\beta }_{\mathit{min}}=0.1$ FA performance per generation: 1	Very high