Parametric variation based simulated annealing for reactive power dispatch

In general power system problems like reactive power dispatch (RPD) are nonlinear, non-differentiable and multi-constrained based. This type of problem is solved by various analytical programming and meta-heuristics algorithms. Among the different applied techniques, metaheuristics algorithms have always maintained themselves competitive with analytical programming. Simulated annealing (SA) is one of the popular meta-heuristics techniques in this applied power system domain due to its simple coding and local-global optimization handling capability. However optimal results of any evolutionary computation techniques including SA depends on the proper choice of the different parameters. In this paper a parametric variation based SA technique is utilized to select optimal parameters for proper convergence. Three case studies with three sets of parameters are exploited to solve RPD problem with IEEE 57-bus system. Among them the best result providing parameter sets are considered as final result. Finally the best result is compared with the other reported results. Comparative study shows the effectiveness of the applied method.