Most of the existing optimization algorithms use fixed size of population. We propose a dynamic population size throughout the optimization process applied on the numerical model of a medium voltage post insulator. Our modified PSO algorithm enables change population in any iteration of optimization process. It is desired to reduce the population size because of a decreasing calculation time. The results of a modified PSO algorithm are compared with a similar modified differential evolution algorithm. Algorithm PSO is suitably modified in order to operate with the principle of the Pareto nondominancy using dynamic population size.