Genetic algorithm for optimization of energy systems: Solution uniqueness, accuracy, Pareto convergence and dimension reduction
Document Type
Article
Publication Date
1-1-2017
Department
Mechanical Engineering
Abstract
Genetic algorithm (GA) is widely accepted in energy systems optimization especially multi objective method. In multi objective method, a set of solutions called Pareto front is obtained. Due to random nature of GA, finding a unique and reproducible result is not an easy task for multi objective problems. Here we discuss the solution uniqueness, accuracy, Pareto convergence, dimension reduction topics and provide quantitative methodologies for the mentioned parameters. Firstly, Pareto frontier goodness and solution accuracy is introduced. Then the convergence of Pareto front is discussed and the related methodology is developed. By comparing two different best points (optimum points) selection method, it is shown that multi objective methods can be reduced to single objective or lower dimensions in objective functions by using ratio method. Our results establish that our proposed method can indeed provide unique solution of satisfactory accuracy and convergence for a multi-objective optimization problem in energy systems. © 2016 Elsevier Ltd
DOI
10.1016/j.energy.2016.12.034
First Page
167
Last Page
177
Publication Title
Energy
Recommended Citation
A. Ganjehkaviri, Mohd Jaafar, M. N., Hosseini, S. E., & Barzegaravval, H. (2017). Genetic algorithm for optimization of energy systems: Solution uniqueness, accuracy, Pareto convergence and dimension reduction. Energy 119: 167-177. doi: 10.1016/j.energy.2016.12.034
Comments
At the time of publication, Seyed Ehsan Hosseini was affiliated with Universiti Teknologi Malaysia.