Sensitivity analysis of GA parameters for ECED problem

Abstract

To meet the requirement of the regulations and to reduce the pollution to the targeted level, minimization of emission level has been added into the dispatch strategies of generators by formulating emission constrained economic dispatch (ECED). Besides the conventional method by using the Lagrange Multiplier several evolutionary computation techniques such as Genetic Algorithm, Particle Swarm Optimisation, Ant Colony and Differential Evolution have been gaining popularity in solving general economic dispatch problems due to their desirable characteristics such as non-gradient dependent and ability to search for global optima. The effectiveness of these stochastic search techniques however is heavily dependent on the genetic operators and their parameters. The paper presents the study on sensitivity analysis of the parameters of Genetic Algorithm (GA) for the ECED problem. The results discuss the range of parameters suitable to be employed for the optimization and compare the difference between conventional economic dispatch and the ECED solutions. © 2013 IEEE.