-
Gustavo M. O. de Castro
eAmazônia - Energia Sustentável e Inovação, AC
-
Aline S. Gallina
eAmazônia - Energia Sustentável e Inovação, AC
-
Moisés Arthur P. Borges
Pós-Graduação em Engenharia da Informação, Universidade Federal do ABC, SP
-
Ronaldo Francisco R. Pereira
Centro de Ciências Exatas e Tecnológicas, Universidade Federal do Acre, AC
-
José Humberto A. Monteiro
Centro de Ciências Exatas e Tecnológicas, Universidade Federal do Acre, AC
Keywords:
Economic Dispatch, Meta-Heuristics, Python, Particle Swarm Optimization
Abstract
This study aims to present the application of the Particle Swarm Optimization (PSO) algorithm to solve the Economic Load Dispatch (ED) problem in its classical formulation, considering losses in transmission lines, and the limits of the generating units. Two case studies were made to demonstrate the method’s applicability, for systems with three and six generating units. Also, a comparison between two different stopping criteria was performed, by the maximum number of iterations and by error. The PSO was implemented using the Python programming language, in the Jupyter Notebook environment. The results were compared against each other, and against the classic optimization solution via Lagrange multipliers present in the literature. The results showed PSO’s applicability for the considered DE problems. Another conclusion of this study is that the stopping criterion by error may return unreliable results since, during the 20 algorithm executions performed in the experiments, this criterion increased the variance of the analyzed metrics and the number of iterations. Therefore, it is concluded that the number of iterations must be known for each experiment, which guarantees the robustness and reliability of the results, considering the smaller standard deviation and percentage error in this case.
