Genetic and genetic/simulated-annealing approaches to economic dispatch

Abstract

The paper first develops and presents the implementation of basic and incremental genetic-algorithm approaches for the determination of the global or near-global optimum solution for the economic dispatch problem. To improve the performances of these algorithms, another algorithm is developed based on the combination of the incremental genetic-algorithm approach and the simulated-annealing technique. This algorithm is then further developed to minimise the memory requirement. A method to overcome the discretisation problem in encoding generator loadings is proposed. A method for ensuring that the dispatch solutions generated in the solution process are feasible and valid is included in the algorithms. The developed algorithms are demonstrated through their applications to determine the economic loadings of 13 generators in a practical power system. In the application study, the effects of valve-point loading and ramping characteristics of the generators are taken into account. The developed algorithms are shown to be general and are computationally faster than the earlier simulated annealing-based method.