Novel approaches for optimal reactive power dispatch using Bio inspired techniques

Abstract

The main objective of this research work is to formulate algorithms basedon bio inspired techniques for solving Multi Objective Optimal Reactive PowerDispatch (MOORPD) problem. Since the ORPD is a combination of discrete andcontinuous variables with multiple local optima, it is difficult to acquire globalvoptima by using classical optimization techniques such as Newton Raphson method,linear programming, quadratic programming and interior point method. In thisthesis, algorithms based on four Bio-inspired techniques such as Firefly, GRADE ,GSO and AGPSO have been formulated to solve the Multi Oobjective OptimalReactive Power Dispatch(MORPD) problem.Firefly algorithm is formulated from a comparative study of PSO based onswarm intelligence which mimic the swarm behaviours such as fish and birdschooling. Based on the three rules of flashing characteristics of fireflies, the Fireflyalgorithm is customised to solve multi objective reactive power dispatch problem.GRADE algorithm is a combination of Genetic Algorithm and DifferentialEvaluation Algorithm. By combining Genetic algorithm and Differential EvaluationAlgorithm, the GRADE algorithm exhibits merits of both GA and DEA for solvingMORPD problem. The proposed GRADE algorithm is found to be capable of bettersolution for MORPD without trapping in local optimum and achieving fastconvergence rate. This is because Genetic algorithm diversifies the particle positioneven though the solution is worse and Differential Evaluation employs efficientpenalty parameter less method for constraints handling.GSO algorithm is primarily suited for solving continuous optimizationproblems such as Multi Objective ORPD. The GSO formulated in this research workis the population based optimization method which employs PS (Producer- Scrounger)model as a frame work. newline