Parallel genetic algorithms for optimization of Modular Neural Networks in pattern recognition

Theory and Background

“…There are many works of different bio-inspired algorithms, but only works of the genetic algorithm and particle swarm optimization and the most important and relevant for this paper will be considered here: Valdez et al [96], performed a hybridization of GA and PSO algorithm and integral the concept of fuzzy logic in it, and other researchers [97,98].…”

Optimization of modular granular neural networks using a hierarchical genetic algorithm based on the database complexity applied to human recognition

“…a method to find an optimal architecture is not used. The parameters of these kinds of neural networks can be the number of hidden layers, the number of neurons for each hidden layer, the learning algorithm, number of epochs, among others [31,32,51]. These kinds of NNs need two phases for their design, which are called the training phase (learning phase) and the testing phase.…”

“…Some of these methods are; the Digital Curvelet Transform using Support Vector Machine (SVM) [48], the decomposition of images into its curvelet sub-bands and applying PCA (Principal Component Analysis) [49], the Spatially Confined Non-Negative Matrix Factorization (SFNMF) method [36], Modular Neural Networks (MNNs) and their optimization using Genetic Algorithms (GAs) or Parallel Genetic Algorithms (PGAs) [51], the use of fuzzy logic as method to perform the combination of responses of modular neural http://dx.doi.org/10.1016/j.ins.2015.02.020 0020-0255/Ó 2015 Elsevier Inc. All rights reserved. networks [28,31], eigenfaces [64], thermal facial patterns using fuzzy neural network techniques [11] or using the similarity measurement of Gaussian maximum likelihood [25] among other techniques.…”

A Fuzzy Rough Rule Based System Enhanced By Fuzzy Cellular Automata

“…Fuzzy Logic is used to combine the results of the PSO and GA in the best way possible. Also, fuzzy logic is used to adjust parameters in the FPSO and FGA [7] [8] . Other research described the use of Modular Neural Networks (MNN) for pattern recognition in parallel using a cluster of computers with a master-slave topology.…”

“…The first neighbor of a cell that changes state, from empty to a class, will result in that its neighborhood cells changing state in the next time step [6] . If 0 means that the cell is empty and the instance's classes are 1 and 2 then the n4V1 nonsatble update rule will look like: (7) where rand (1, 2) selects randomly from the elements with equal probability.…”