Hybrid Optimization Algorithm for Reactive Power Optimization with Discrete Variables

Hybrid optimization methods are used to resolve reactive power optimization with discrete and continuous variables,by which the complementary characteristics of genetic and traditional optimization algorithms are utilized.Genetic operations(such as selection,crossover and mutation)are acted on discrete variables only.The genetic algorithms are used to make the global searches to the population.Before a new offspring is selected into the population,it must move to the local optimal point by using the traditional optimization methods to the continuous variables.To ensure the effects of the local optimization,a new optimization algorithm which is based on function transform and generalized inverse of matrices is used.The model of hybrid optimization algorithm is simplified and normal,which has the advantage of genetic algorithms which treated discrete variables conveniently and traditional optimization methods which calculated fast and steady.The practicability and effectiveness of algorithms are proved by case studies.