Abstract
Methods of direct search are evolutionary algorithms used for constrained optimization problem solving. Information regarding objective function gradient is not required for finding an optimum solution. Pattern search is one such method. This paper gives an application of the method for non-convex economic dispatch problem solving. As a confirmation of its effectiveness, the method has been tested on the systems consisted of five, thirteen and forty thermal power units. Also, features of this method have been compared to other popular methods. The results are encouraging and show that the pattern search method is a very useful for non-convex economic dispatch problem solving.
© 2016 Damir Aganović, published by Bosnia and Herzegovina National Committee CIGRÉ
This work is licensed under the Creative Commons Attribution 4.0 License.