Proposição e implementação de um controlador de cargas inteligente considerando custo energético e conforto baseado em algoritmos genéticos

Abstract

There is a demand for optimization of inefficient processes that generate additional costs. The present work presents a proposition and implementation of a code, through a charge controller, based on genetic algorithm. The selection of this evolutionary algorithm is due to the need of resolving a multiobjective optimization problem, in order to maximize economy and comfort through relocation of charges in a house. The approach used focus on the analysis of time of code execution and the solutions’ resulting hypervolume indicators. This hypervolume is generated through Pareto’s Front, obtained by the distances between the solutions and what is considered a great point. A scenery of charges in which the times can be modified was used to compose the code, the first generation being made of previously selected scenarios to facilitate the convergence. These scenarios have their skills tested by the goal functions for economy and comfort. The Pareto Envelope-based Selection Algorithm II (PESA II) is utilized to generate Pareto’s Front. The code is executed in various generations settings, and several times, with the intent to find the best answers between those scenarios. The simulations results show great evolution on the increasing of generations, granting more stability and better results. What sets the methodology apart is the better understanding of the impact of the algorithm’s inputs and outputs and its performance, since it’s a code in search of solutions, with many randomness elements and dependent of the variety of its solutions.