The aim of this paper is to introduce a biologically inspired approach that can automatically generate Deep Neural networks with good prediction capacity, smaller error and large tolerance to noises. In order to do this, three biological paradigms were used: Genetic Algorithm (GA), Lindenmayer System and Neural Networks (DNNs). The final sections of the paper present some experiments aimed at investigating the possibilities of the method in the forecast the price of energy in the Brazilian market. The proposed model considers a multi-step ahead price prediction (12, 24, and 36 weeks ahead). The results for MLP and LSTM networks show a good ability to predict peaks and satisfactory accuracy according to error measures comparing with other methods.
Classical models of agents for solving well-defined problems are widely used in the literature but are limited to systematic search strategies in order to find the solutions. However, these strategies are not suited for all types of application. This work presents an adaption of classical models of agents for local search strategies. One agent system for neural network automatic design is used to show the feasibility of the proposal. The results are promising, since the model found satisfactory solutions for the proposed problems.
This paper proposes a hybrid approach for shortterm energy price prediction. This approach combines a genetic algorithm (GA) that evolves individuals represented by a set of production rules, these rules are extracted from chromosomes and generate the genotypes, which are mapped to phenotypes(Deep Neural Networks). The Artificial Neural Networks (ANNs) are then trained and then validated, and the prediction tests are carried out. The genotypes are classified by the performances of their ANNs in prediction and the GA selects the best individuals for mutation and crossover operations, which provide a new population. The previous steps are repeated through n generations. The result is an optimized neural network architecture for energy price prediction. The results show good ability to predict spikes and satisfactory accuracy according to error measures, delivering an accurate prediction. Finally, the results are compared with traditional techniques.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.