In microwave hyperthermia tumor therapy, optimization of specific absorption rate (SAR), is the conventional approach to focus the heat energy into the target region. Two versions of particle swarm optimization (PSO) were investigated, iw-PSO and CL-PSO, as the optimization techniques for the antenna array excitations in the quest. A two-layer 16-antenna fractal octagonal ring antenna (FORA) circular array was used as the MH applicator. The array lattice possesses symmetry infrastructure for which it is subdivided into iterative scaled-down sequences of the outer octagonal ring. The optimization was conducted on a 3D realistic breast model. The general parameters of the PSO algorithm were kept the same for the two versions and the results were compared in terms of target-to-breast temperature ratio and the average temperature at the healthy tissue. The results demonstrate that CL-PSO provided a remarkable performance over iw-PSO with the same number of iterations and population size.