ABSTRACT
Reactive power compensation at per level is one of the fundamental operational issues of power systems operation and control. To solve the reactive power compensation problem, the conventional real coded genetic algorithm (RCGA) is customized by the concept of self-adaptation of multivariate q-Gaussian distribution. This is followed by a polynomial mutation approach with arithmetic crossover for optimal control of reactive power in distribution feeders to evade premature convergence with fast rate of convergence. The improved technique has been added to the conventional RCGA. A new combination of crossover and mutation technique has been adopted to get a better route for escaping the local minimum solution. The performance of the algorithm is verified on a sample test system and the better simulation results turn out improved solutions compared to different methods.
